Effects of anthropogenic disturbance on tree population
structure and diversity of a rain forest biosphere reserve in
Ghana, West Africa
Emmanuel Morgan Attua1* , Louis Awanyo2 and Effah Kwabena Antwi3
1Geography and Resource Development, University of Ghana, P. O. Box LG 59 Legon Accra, Ghana, 2Luther College, University of Regina, Regina,
Saskatchewan, SK S4S 0A2, Canada and 3Institute for the Advanced Study of Sustainability, United Nations University, Tokyo, Japan
Abstract R
esum
e
We evaluated the impacts of anthropogenic disturbance Nous avons 
evalu
e les impacts des perturbations anthro-
on community structure and diversity along three pogenes sur la structure et la diversit
e de la communaut
e
management zones of the Bia biosphere reserve in v
eg
etale dans trois zones de gestion de la R
eserve de
Ghana. Sixty sample plots were distributed among the Biosphere de Bia, au Ghana. Soixante parcelles 
echan-
core, buffer and transition zones. We estimated the tillons ont 
et
e r
eparties dans les zones centrale, tampon et
degree of disturbances from discernible indicators on the de transition. Nous avons estim
e le degr
e de perturbation
field and satellite images. All tree species ≥10 cm dbh au moyen d’indicateurs et d’images satellitaires. Toutes les
(diameter at breast height) were identified and enumer- especes d’arbres d’un dbh ≥ a 10 cm ont 
et
e identifi
ees et
ated. Inventory data were compared across the zones d
enombr
ees. Les donn
ees des inventaires ont 
et
e com-
and related to intensity of disturbances. A total of 1176 par
ees entre les zones, et li
ees a l’intensit
e des perturba-
individual trees from 108 species and 33 families were tions. Nous avons compt
e un total de 1 176 arbres de
encountered. Number of species varied from 27 in the 108 especes et 33 familles. Le nombre d’especes allait de
highly disturbed (HD) to 61 in the least disturbed (LD) 27 dans la zone la plus perturb
ee (HD) a 61 dans la moins
zone. Mean basal area (BA) varied from 11.71 in the HD perturb
ee (LD). La surface basale moyenne variait de
to 28.26 in the LD. Both Margalef’s species richness and 11,71 dans la zone HD a 28,26 dans la zone LD. Et l’indice
Shannon-Weiner’s a-diversity were highest in the mod- de diversit
e de Margalef et la diversit
e-a de Shannon-
erately disturbed (MD) than either the least and most Wiener 
etaient plus 
elev
es dans les zones mod
er
ement
disturbed zones. Our study revealed significant differ- perturb
ees que dans celles qui l’
etaient plus, ou moins.
ences in tree abundance, stem density, BA and species Notre 
etude a r
ev
el
e des diff
erences significatives dans
diversity, attributable to differences in degree of anthro- l’abondance des arbres, la densit
e des troncs, la surface
pogenic disturbances among zones. Given the different basale et la diversit
e des especes, que l’on peut attribuer
levels of anthropogenic disturbance and corresponding aux diff
erences de perturbations anthropogenes entre les
impacts across the reserve, we recommend an integrated zones. E
tant donn
e ces diff
erences de perturbations et les
management strategy for the conservation of biodiversity impacts correspondants dans toute la r
eserve, nous
in the Bia biosphere reserve. recommandons une strat
egie de gestion int
egr
ee pour la
conservation de la biodiversit
e dans la R
eserve de
Key words: anthropogenic disturbances, Bia biosphere
Biosphere de Bia.
reserve, Ghana, intermediate disturbance hypothesis, tree
structure, tropical forests
Introduction
Prior to forest management activities, quantifying bio-
physical attributes, such as structure, species composition
*Correspondence: E-mail: emattua@ug.edu.gh and diversity in response to disturbances, offers relevant
116 © 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
Effect of disturbance in Bia biosphere reserve 117
information for restoring degraded landscapes (Defries 2000) and also a World Heritage Site for UNESCO
et al., 2005; Hansen & Defries, 2007). Rain forests hold the (Forestry Commission, 2010). Besides, it is the first
highest levels of species richness and endemism of any biosphere reserve in Ghana, designated in 1983 by
terrestrial biome (MEA, 2005). They are, therefore, usually UNESCO. The reserve is therefore of both national and
the target of most biodiversity conservation initiatives global conservation significance. Unfortunately, the latest
(IUCN, 2010). In Africa, huge areas of rain forests and assessment report on the management effectiveness of
their resident biodiversity are decimated annually (FAO, parks and reserves in Ghana revealed that the biosphere
2010; Gardner et al., 2010). This is on account of human reserve was experiencing significant disturbances, largely
disturbances through mainly agricultural conversions, from logging and agricultural activities (UICN/PACO,
lumbering, settlement development, road constructions 2010). The need for effective management by the Wildlife
and extraction of nontimber forest products (Malhi et al., Division of the Forestry Commission, to conserve the
2013). Despite the high disturbances, most protected rain unique habitat and biodiversity of the biosphere reserve, is
forests conserve indigenous biodiversity better than other long overdue (Ministry of Environment and Science,
land uses on the continent (Struhsaker, Struhsaker & Siex, 2002).
2005). It is evident therefore, more than ever that the Although not regarded as formal protected areas,
long-term sustainability of existing African rain forest biosphere reserves have their origin from the protected
reserves is greatly dependent on how disturbances are area concept (Ishwaran, Persic & Tri, 2008; IUCN & UNEP-
understood and effectively managed in protected and WCMC, 2014). Unlike ‘classical’ protected areas, however,
adjacent landscapes (Harvey et al., 2008; Wittemyer et al., the overall management of biosphere reserves encom-
2008; Gardner et al., 2009). passes both protected and nonprotected areas (Hansen &
For many African countries, protected areas are essen- Defries, 2007). Lessons from study of biosphere reserves,
tial elements of major intervention strategies to conserve therefore, have implications that go beyond the particular
biodiversity and, probably, the only means of conserving study area and can be applied to both managed and
obligate forest species (Gardner et al., 2009). An enhanced unmanaged landscapes (Batisse, 1993; German Commis-
understanding of disturbance regimes and impacts on sion for UNESCO, 2015). In this study, we sought to
biodiversity is thus vital for the sustained management of quantify the degree of disturbances and impacts on stand
these protected rain forests. Particularly in Ghana, structure, tree composition and diversity in the Bia
although some studies are available on the subject (e.g. biosphere reserve. Specifically, we sought to: (i) compar-
Hawthorne, 1993; Hall et al., 2003; Addo-Fordjour et al., atively analyse the degree of disturbances across the
2009; Pappoe et al., 2010), there is little agreement in management zones of the reserve; (ii) determine the
knowledge for generalized application. This is for a number relationship between disturbance and forest stand struc-
of reasons, including over-focusing on single disturbance ture, tree species composition and alpha diversity between
events such as selective logging and fire (Hawthorne, 1993; zones; and (iii) proffer appropriate strategies on basis of
Hall et al., 2003; Bongers et al., 2009; Wiafe, 2014) and findings from (i) and (ii), for better management of tree
differences in study scale and methodology. Moreover, biodiversity in the reserve. In concurrence with the
notwithstanding the several studies on effect of disturbance intermediate disturbance hypothesis, we posited that
on species diversity, there is no consensus yet on the extreme anthropogenic disturbances in forest biosphere
diversity disturbance debate; thus requiring further work to reserves eliminate sensitive tree species, but moderate
explicate. For example, according to the refugia hypothesis, levels of disturbances encourage more species to co-exist
undisturbed habitats are of higher diversity compared to leading to high species richness and diversity in less
disturbed habitats because they serve as refugia for disturbed areas.
organisms (Townsend, 1989; Townsend & Hildrew,
1994). On the contrary, the intermediate disturbance
Materials and methods
hypothesis postulated by Connell (1978) suggests that
diversity within communities is maximal at intermediate
Study area description
frequencies and intensities of disturbances.
The Bia biosphere reserve of Southwestern Ghana is The Bia biosphere reserve is situated in Southwestern
among the world’s 25 biodiversity hotspots (Myers et al., Ghana, only about 5 km to the border with La Côte
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
118 Emmanuel Morgan Attua et al.
Fig 1 Location map of the Bia biosphere
reserve in context of the Western Region of
Ghana, showing management zones and
road network linking major communities.
[Colour figure can be viewed at wileyonli-
nelibrary.com]
d’Ivoire. Lying between latitudes 06°200–06°390N, and Abu-Juam, 1995). On the other hand, the vegetation of
longitudes 02°580–3°130W (Fig. 1), the reserve covers an the core (National Park) is generally undisturbed forest.
area of 355.6 km2. It comprises a core area (also a The Bia biosphere reserve forms part of the vulnerable
National Park) of 77.7 km2 to the north, a buffer area upper Guinean rain forest, a strip of tropical forest that
(also a Resource Reserve) of 227.9 km2 to the south and a stretches from Sierra Leone to Ghana (Forestry Commis-
transition zone of 837 km2. The transition area is dom- sion, 2010). It lies between the moist evergreen and semi-
inated by large cocoa farms strewn with few trees and deciduous humid tropical forest of Ghana. The original
dotted with food crop farms (Forestry Commission, 2010). vegetation was therefore characterized by tree species of
The area now corresponding to the buffer zone was excised both forest types. These included Lophira alata and
from the National Park and opened up for timber Scaphopetalum amoenum which are characteristic species
exploitation until it was constituted as a Game Production of the evergreen forest; and Khaya ivorensis, Khaya
Reserve in 1983 (Hawthorne et al., 2001). The vegetation anthotheca and Entandrophragma utile which are unique
of the buffer (Resource Reserve) was disturbed by logging to the semi-deciduous forest (Hall & Swaine, 1981). The
activities and wildfire up to 30 years ago (Hawthorne & reserve is also home to some 404 species of butterflies, 130
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
Effect of disturbance in Bia biosphere reserve 119
species of birds, the African elephant, chimpanzees and
other primates (UICN/PACO, 2010).
The reserve lies in the wet semi-equatorial climatic zone
and experiences alternating wet and dry seasons. The dry
period is from December to March. Annual temperatures
vary from a minimum of 20.5–22°C (February/March) to
a maximum of 29–34°C (July/August), with mean tem-
peratures between 24°C and 28°C (Forestry Commission,
2010). The area experiences a bimodal annual rainfall
regime with two seasonal peaks in June and September.
The annual mean precipitation is about 1500 mm. The
topography is undulating and generally flat with elevation
ranging between 168 and 238 m. The geology is mixed,
with Lower Birimian to the east, granites to the west and Fig 2 Species–area curves for all sample trees ≥10 cm dbh in the
Upper Birimian forming a north–south strip through the Bia biosphere reserve, Ghana
middle. The soils are acrisols, locally classified as forest
ochrosols and are generally of red or reddish brown species density and richness are both not linearly related to
appearance. These soils are of moderate acidic reaction, plot size, we did not adjust calculations for size, but all data
with pH of between 6 and 7 (Dickson & Benneh, 1998). were used directly. Species were identified on the field and
The human population is settled by over 40 communities, where this was impossible, botanical specimens were
all in the transition zone and is largely agrarian, with collected, taken to the National Herbarium at the Univer-
majority being cocoa and subsistence food crop farmers. sity of Ghana, for later identification. The names of all
species were standardized using the Taxonomic Name
Resolution Service (tnrs.iplantcollaborative.org/
Sampling and data collection
TNRSapp.html), the Plant List (www. theplantlist.org)
The sampling sites were the core, buffer and transition and the literature (Taylor, 1952; Hutchison et al., 1957-
zones of the biosphere reserve. The field team (including 1972; Hawthorne, 1990). All nomenclature followed the
the Principal author) used a sampling procedure of Hall & International Plant Nomenclature Index (IPNI, 2008).
Swaine (1981), as modified by Hawthorne & Abu-Juam
(1995). In brief, twenty sampling plots each
Estimation of disturbance levels and index
30 m 9 50 m dimension were randomly distributed along
three near-parallel 1 km transects in each site, avoiding The three management zones of the biosphere reserve
only inaccessible areas such as highlands and large streams. were sampled separately. They were ranked according to
Thus, in total 9 ha was sampled across the biosphere the degree of anthropogenic disturbance, using a hybrid
reserve. According to Lamprecht (1989), theminimumarea method introduced by Sagar, Raghubanshi & Singh
needed to sufficiently sample a plant community is the point (2003) and Kumar & Shahabuddin (2005). Prior to tree
when increment in number of species levels off to below10% inventorying, all major disturbance regimes and their
of total species. About 3 ha was considered the minimum severity were discussed with managers of the reserve and
area for adequate enumeration of tree species at each site, as selected farmers. From satellite images, we were able to
evident from species–area curves (Fig. 2). estimate three indicators of disturbance: degree of agri-
On each plot, we tagged and measured diameter of all cultural activities (represented as proportion of cultivated
trees ≥10 cm at breast height (dbh) over-bark and at area to total area), proportion of human settlements to
1.37 m above ground and identified to species. Diameter of total area and number of roads/footpaths in each zone.
trees with buttresses was measured 50 cm above the Four additional indicators were assessed on sampling
buttresses. Forked trees were recorded as separate individ- plots, namely proportion of trees cut or lopped (from
uals if branching occurred below 1.37 m above ground logging and other wood harvesting activities), degree of
otherwise were counted as single individuals. Because the lopping (on a scale of 0–5 for each tree: 0 = no lopping;
same sampling approach was used on all plots and as 1 = rudimentary signs of lopping; 2 = up to a quarter of
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
120 Emmanuel Morgan Attua et al.
Table 1 Estimated relative disturbance intensities in the Bia standard formulae within PAST 3 (a PAlaeontological
biosphere reserve of Southwestern Ghana Statistics) software (Hammer, Harper & Yan, 2001) as
follows (Kent & Coker, 2002):
Relative impacts of
disturbance
S 1
Indicators of disturbance LD MD HD SR ¼ ð Þ (1)ln N
Agricultural activities 0 0 5
Human settlements 0 0 5 ¼ SEw  (2)Cutting/Lopping 0 1 5 lnNi lnNs
Tree canopy openness 1 2 5
Ground vegetation cover 1 2 4 s0 X
Disturbed soil/bare 1 2 4 H ¼ pi ln pi (3)
i¼1
Footpaths/skid trails/roads 1 2 4
Total 4 9 32  X !s
D ¼ 1 p2i : (4)
i¼1
main branches lopped; 3 = up to half of main branches
lopped; 4 = more than half of main branches lopped; and Whittaker’s b-diversity index (bw) was calculated based
5 = nearly all main branches lopped or tree reduced to a on the twenty 30 m 9 50 m sample plots of each site. The
stump by lopping or harvesting; the total lopping score values, therefore, reflect habitat diversity within a zone
was then divided by the total number of trees present), and were calculated as (Whittaker 1972; cited in Kent &
extent of tree canopy openness, proportion of ground Coker, 2002):
vegetation cover and per cent bare ground/disturbed soil
cover. We calculated the relative impacts of disturbance ¼ Sbw : (5)
in each management zone by finding the total score of all Ŝ
plots. As per the computations, the core, buffer and
transition zones were designated as least disturbed (LD), In Eqs 1–5 above, S = total number of species; N = total
moderately disturbed (MD) and highly disturbed (HD), number of individuals of all species; ni = number of
hereafter referred to as LD, MD and HD respectively individuals of the ith species; pi = ni/N = the proportion
(Table 1). of individuals belonging to the ith species; Ni = number
Also, for each sample plot, a disturbance index (DI) was of individuals of most important species; Ŝ = average
calculated as the percentage of trees in the plot that are number of species per sample plot; ln = natural log (i.e.
pioneers (Bongers et al., 2009), taking into account all 2.718). We also estimated stand structure of each site on
pioneer species ≥10 cm dbh. This index was used because basis of BA and diameter class distribution data. Basal
it reflects disturbance history at different scales in Ghana- area of an individual tree was estimated as (Kent &
ian forests (Bongers et al., 2009). Coker, 2002):  
dbh 2
BA ¼ pr2 ¼ 3:14x (6)
Data analysis 200
Adequacy of sampling was determined from constructed where BA (in m2) is BA, and r is radius of tree (in cm).
species–area curves (Lamprecht, 1989; Brose, Martinez & For analysis, we used one-way analysis of variance
Williams, 2003; Gotelli & Chao, 2013). Following, we (ANOVA) to establish significant differences in tree
analysed community composition at species, genera and density and BA among the three disturbance zones.
family levels, using inventory data. Species richness Also, we grouped all sampled trees into four diameter
(number of species per sample plot), evenness (distribution classes, following Swaine and Hall (1988): poles (10–
of abundances among species) and a-diversity (level of 19.9 cm dbh), small trees (20–39.9 cm dbh), medium
species heterogeneity) were calculated using Margalef’s trees (40–69.9 cm dbh) and large trees (≥70 cm dbh).
index (SR), Shannon-Weiner a-diversity index (H0) and For each diameter class, we calculated stem density and
Simpson’s a-diversity index (D). These were computed from BA and subjected the results to ANOVA to determine
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
Effect of disturbance in Bia biosphere reserve 121
significant differences in these values among the diam- genera and 28 families) and the least in HD (149 individ-
eter classes. We applied Tukey’s honest significant uals; 42 species – from 27 genera and 18 families). Themost
difference (HSD) comparison test at a 5% probability abundant species encountered belonged to the Fabaceae
significance threshold to compare all means that exhib- (seventeen species), Malvaceae (fourteen species) and Meli-
ited differences. All data used for statistical analyses aceae (thirteen species) families. Among the zones, species
were tested for normality and homogeneity of variance belonging to the Fabaceae familywere preponderant in both
assumptions, and where necessary appropriate data LD and MD (fifteen species each). However, in HD, the
transformations were carried out. Species richness data Malvaceae were the most represented (seven species)
were square-root transformed to meet the above-men- followed by Fabaceae (six species) and Moraceae (five
tioned assumptions. Nonetheless, data on tree abun- species). Number of species per study plot ranged from 42
dances and stem density could not be normalized, and, inHD to 102 inMD. Similarly, number of genera varied from
therefore, non-parametric Kruskal–Wallis test (H) test 27 in HD to 91 in MD. A number of unique species were
was used to compare their differences among zones. identical between sites; being between 4 and 5. A number of
Significance level was set at P = 0.05 (two-tailed) for all species with more than ten individual trees were few and
tests. Relationship between level of disturbance and ranged between 1 and 3, indicating codominance of species
stand structure was determined by running Pearson’s in the vegetation (Table 2).
correlation analysis between the following parameters: Overall, Ceiba pentandra, Antiaris toxicaria, Terminalia
(i) DI versus BA and (ii) DI versus square-root of species superba, Erythroxylum mannii and Celtis mildbraedii were
richness (√SR). the five most abundant species in all zones (see Table S1).
The three most abundant tree species in the LD zone were
C. pentandra, A. toxicaria and E. mannii, constituting
Results
10.19% of total individuals. In the MD forest, C. pentandra,
A. toxicaria and Nesogordonia papaverifera were the three
Effects of disturbance on tree abundance
most abundant species, constituting 10.25% of all trees in
Across the three zones, inventories yielded a total of 1176 the zone. In the HD zone, Elaeis guineensis, Mangifera indica
individual trees comprising 108 species belonging to 91 and C. pentandra were the three most abundant tree
genera and 33 families (Table S1). The highest number of species, making up 26.85% of all standing trees in the
individual trees and species was encountered in MD (556 HD part of the biosphere.
individuals; 102 species – from 91 genera and 32 families) The most disturbed zone had significantly fewer trees
followed by LD (471 individuals; 87 species – from 77 than either the least or MD zones. A Kruskal–Wallis test
Table 2 Pattern of tree stand composition and structure in Bia biosphere reserve grouped by level of disturbance
LD MD HD
a. Stand composition and structure
Number of species 87 102 42
Number of genera 77 91 27
Number of families 28 32 18
Number of unique species 4 5 4
Number of species with >10 individuals 3 3 1
Mean stem density (≥10 cm dbh) 107  49 171  53 69  29
Mean basal area (m2 plot1) 28  7.70 21  7.94 11  9.62
b. Species richness and diversity
Species richness (Margalef’s index) 11  0.72 13  3.91 4  2.62
Evenness (Whittaker’s index) 4.73  1.28 4.67  1.14 1.85  0.72
Shannon-Weiner’s a-diversity 4.22  0.63 4.61  1.15 1.08  0.87
Simpson’s a-diversity (1-D) 0.94  0.25 0.97  0.21 0.64  0.13
Whittaker’s b-diversity 1.69 1.71 1.72
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
122 Emmanuel Morgan Attua et al.
(H) revealed that there was a significant effect of anthro-
pogenic disturbances on tree abundances in the three
zones of the biosphere reserve (H = 45.19; v2 = 5.99;
df = 2; P < 0.05). The order of tree abundance among the
zones was: HD ˂ LD ˂ MD).
Effects of disturbance on species richness and diversity
Generally, increasing disturbance decreased evenness in
distribution of species (Table 2).Whittaker’s evenness index
was identical in LD and MD but was very low in HD. Both
species richness and a-diversity (both Shannon-Weiner’s
and Simpson’s) were highest at intermediate disturbance.
Species richness was highest in MD and lowest in the HD. A Fig. 3 Distribution of stem diameter sizes by levels of disturbance
similar trend was seen for both Shannon-Weiner’s a- in the Bia biosphere reserve of Southwestern Ghana
diversity and Simpson’s a-diversity. Surprisingly, b-diversity
was identical across all sites and seems not to respond to between MD and HD but not between LD and MD (Tukey
degree of anthropogenic disturbance. HSD; P = 0.117). Differences in density of large trees
between LD and HD and between MD and HD were both
significant. Nonetheless, no significant difference was
Effect of disturbance on stand structure
found between LD and MD (Tukey HSD; P = 0.381).
Due to differences in disturbance, mean BA varied from Generally, numbers of both medium and large trees
11.71 m2 ha1 in HD to 28.26 m2 plot1 in LD (Table 2). declined with increasing level of disturbance.
There were significant differences in BA among all the Generally, all diameter distributions were positively
pairs of the zones (F2,3 = 3.27, P = 0.002). For each pair, skewed (Fig. 3) towards the small diameter cohort.
the less disturbed zone had significantly higher BA than However, tree population by diameter size varied greatly
the more disturbed zone (i.e. LD > MD, LD > HD, between the zones. In HD, nearly 75% of the trees were
MD > HD). BA of trees in MD and HD was 24% and either poles or small trees compared to about 23%
59% less than that of LD respectively. Large standard medium and 2% large trees, none above 80 cm dbh. In
errors in variables suggest marked stand variability MD, about 58% trees were either poles or small trees,
between the zones. Stem density of tree ≥10 cm dbh 35% medium trees and 7% large trees. In LD, about
differed across zones with mean ranging between 69 in HD 47% trees were either poles or small trees, 37% medium-
and 171 in MD (Table 2). There were significant differ- size trees and 16% large trees. Broadly, increasing
ences in stem density (H = 14.92; v2 = 5.99; df = 2; disturbance corresponded to decreasing population of
P < 0.05) among all the pairs of zones due to anthro- large-size trees.
pogenic disturbances in the reserve. For each pair, the MD A negative linear relationship was also observed when
zone recorded significantly higher stem density than the DI was regressed on BA and species richness (Fig. 4). Thus,
LD zone (i.e. MD > LD, MD > HD and LD > HD). increasing forest disturbance consistently led to a decline
The density of all diameter classes significantly varied in BA and species richness in all zones. Comparatively, the
between zones: poles (F2,3 = 5.02; P < 0.001); small trees strength of relationships (R
2) was in the order: LD >
(F2,3 = 7.52; P < 0.004), medium trees (F2,3 = 11.41; MD > HD.
P < 0.009) and large trees (F2,3 = 6.83; P < 0.000). At
P < 0.05 significance, Turkey’s test revealed that there
Discussion
were significant differences in density of both poles and
small trees, between LD and HD and between LD and MD.
Effect of disturbance on tree species richness and diversity
Nonetheless, the difference between HD and MD was
insignificant (Tukey HSD; P = 0.085). Density of medium Local species richness and diversity can be markedly
trees differed significantly between LD and HD and influenced by disturbances, and depending on type of
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
Effect of disturbance in Bia biosphere reserve 123
5 30
(a) but declined with increasing anthropogenic disturbance.
4 25 Our observation concurs with other studies in other
tropical forests outside Africa (Sheil, 1999; Zhu et al.,
3 20 2007) and in Ghana as well (Addo-Fordjour et al., 2009).
2 15 Our findings also lend credence to the intermediate
1 disturbance hypothesis which posits that extreme distur-BA = –0.0797DI + 5.7678; R2 = 0.7192; P = 0.00 10
SR = –0.5807DI + 35.189; R2 = 0.7067; P = 0.00 bances usually eliminate sensitive species, but moderate
0 5
15 20 25 30 35 40 45 levels of disturbances allow more species to co-exist
Disturbance index, DI (%) (Tilman, 1982; Sheil, 1999; Wilkinson, 1999).
Basal area species richness Our study also documented that disturbance caused a
5 20
(b) significant decline in tree species diversity as revealed by18
16 Shannon-Weiner and Shannon indices. Nonetheless, a
4 14 study of the effects of disturbance on tree species diversity
12
10 along a wet–dry climatic gradient in Ghana’s forests
8
3 revealed that disturbance contributed significantly to the6
2 4 maintenance of tree species diversity in dry forest butBA = –0.3636DI + 25.67; R  = 0.6759; P = 0.00
SR = –0.0924DI + 6.6401; R2 = 0.635; P = 0.003 2 played only a minor role in moist and wet forests
2 0
20 25 30 35 40 45 (Bongers et al., 2009). A combination of factors related to
Disturbance index, DI (%)
local differences in land-use intensification, biodiversity
Species richness (SR) Basal area (BA)
management and on-site history (Hawthorne, 1993;
4 12
(c) Forestry Commission, 2010) as well as study scale
10
3 (Bongers et al., 2009) could be responsible for the
8 discordant outcomes. Indeed, in most tropical regions,
2 6 conversion of forests to farmlands involves the removal of
4 a substantial number of forest trees (Makana & Thomas,
1 BA = –0.0403DI + 4.5194; R2 = 0.3958; P = 0.007
2 2 2006). Agricultural land-use intensification, usuallySR = –0.139DI + 13.397; R  = 0.495; P = 0.001
0 involving slash-and-burn cultivation, is rather pervasive0
35 40 45 50 55 60 65 and a major cause of tree species loss in Ghana’s
Disturbance index, DI (%)
Species richness Basal area agroecological zones (Antwi et al., 2014). In the Bia
biosphere reserve, differences in approach to biodiversity
Fig 4 Relationship between disturbance index and species rich- management in the different zones might have exacer-
ness and basal area at the different management zones of the Bia
bated tree species loss as observed in some protected
biosphere reserve, Ghana; (n=20): (a) LD (core zone); (b) MD
tropical rain forests (Defries et al., 2005; Hansen &
(buffer zone); and (c) HD (transition zone)
Defries, 2007).
forests and extent of disturbance, may vary significantly
Effects of disturbance on tree stand structure
(Hall et al., 2012). Our study estimated the lowest species
richness in the HD portion of the biosphere reserve, Our evidence that mean densities of trees varied signifi-
corresponding to the transition zone. This could be cantly between sites indicates high spatial variability in
attributed to frequent disturbances from human activities, tree distribution in the biosphere reserve. Our observation
such as farming, logging and settlement development, concurs with many other studies elsewhere in the tropical
known to occur in nearly all agroecological zones of region (Whitmore & Burslem, 1998; Burslem & Whitmore,
Ghana (Antwi et al., 2014) and especially outside pro- 1999; Sheil, 1999; Ramirez-Marcial, Gonzalez-Espinosa &
tected forests (Ardayfio-Schandorf et al., 2007; UICN/ Williams-Linera, 2001; Sagar, Raghubanshi & Singh,
PACO, 2010). 2003; Sahu, Sagar & Singh, 2008).
In this study, a linear relationship was found between DI This study found a higher number of unique species in
and species richness at all sites. Generally, species richness the HD zone of the biosphere reserve. This can be
was maximum when forest disturbance was only moderate attributed to introduction of species through cultivation
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
√Species richness, SR √Species richness, SR √Species richness, SR
Basal area, BA (m2 per plot)
Basal area, BA (m2 per plot)
Basal area, BA (m2 per plot)
124 Emmanuel Morgan Attua et al.
or invasion by opportunistic species (Hawthorne, 1993; unlikely to succeed, especially under conditions of increas-
Addo-Fordjour et al., 2009). Another key observation was ing land-use pressure (Defries et al., 2005). Selective
that many species found in the LD core and MD buffer logging as a management strategy has been applied to
zones of the biosphere reserve were absent in the HD increase species richness and diversity (Ola-Adams, 1987;
transition zone. According to Lamprecht (1989), high Hawthorne, 1993; Verburg & Van Eijk-Bos, 2003;
utilization pressure could be responsible. Generally, in Edwards & Lawrance, 2013). This may be encouraged in
many forests of Ghana, adult population of many eco- the buffer zone to enhance light penetration for natural
nomic trees is of low densities; averaging one tree per regeneration of pioneer and light-demanding obligate
hectare (Hawthorne & Abu-Juam, 1995; Lauma-Aho, species, especially rare and endemic species (van der Werff
2003; Appiah, 2013), sometimes even less than one & Consiglio, 2004).
commercial tree per 10 ha (Lamprecht (1989). With rapid In the HD transition zone, agroforestry practices on
increases in human population, recurrent activities of and around farms as an alternative to more extensive
shifting cultivation, timber extraction and fuelwood har- and less sustainable land-use practices may improve
vesting could alter the habitat suitability for most tree biodiversity conservation (Garrity, 2004; Makundi &
species (Hawthorne, 1993). Sathaye, 2004; Schroth et al., 2004). This can also help
Usually, ecologists interpret large numbers of juveniles minimize the quest of local people to deforest more land
relative to adults as indicative of a stable community, for agriculture, in addition to coping with limited
perhaps growing; but a few juveniles relative to adults availability of forest land and resources (Schroth et al.,
depict a declining population (Condit et al., 1998). The 2004). Enrichment planting (Lamprecht, 1989; Doucet
particularly high population of poles and small trees in the et al., 2009) could be encouraged as well to replace lost
transition zone suggests active tree regeneration (Hall species. However, given the inappropriateness of exotic
et al., 2012), probably, facilitated by germination and species for replanting degraded lands in Ghana (McCul-
development of pioneer species because of enhanced light lough, Decher & Kpelle, 2005), selective use of native
penetration (Hawthorne et al., 2001). However, the fact species will be preferred to exotic species (Appiah, 2011).
that both species richness and diversity were low in this Selective use of native species will be in concordance with
zone suggests that any biodiversity intervention strategy the Forestry and Wildlife policy of Ghana (Ministry of
should aim at replacing lost species. Land and Forestry, 1994).
The realization of any of the above biodiversity man-
agement objectives in the Bia biosphere reserve will entail
Management options for biodiversity conservation
respecting the ecological integrity of the core conservation
Evidently, anthropogenic disturbance has disproportion- area and managing the environment within which
ately influenced tree species composition, structure and adjoining areas are established for biodiversity conserva-
diversity pattern in the various management zones of the tion. This will ensure that the fundamental principles of
Bia biosphere reserve. Considering that tree species com- the biosphere reserve concept remain honoured. Also,
position, structure and diversity varied greatly along the strengthening any existing community-based forest man-
different levels of forest disturbance, an integrated man- agement structures and promotion of alternative livelihood
agement strategy will be required to achieve effective programmes could help maintain forest structure and tree
biodiversity conservation in the biosphere reserve. With species richness in the long term.
careful application, likely restorative actions could include
(i) protection of intact native forests; (ii) tree canopy
Conclusion
management; (iii) natural regeneration; (iv) enrichment
planting; and (v) agroforestry. This study has demonstrated that forest structure, tree
Our study reveals that extreme anthropogenic distur- species richness and diversity are all influenced along a
bances led to loss of species diversity in the Bia biosphere disturbance gradient. The management zones of the
reserve, particularly in the transition zone. It is therefore biosphere reserve experienced variability in floristics in
reasonable that natural forest protection remains an response to disturbances. An integrated biodiversity man-
integral part of the conservation strategy of the biosphere agement is therefore recommended for rehabilitation of
reserve. Notwithstanding, forest protection alone is disturbed areas of the biosphere reserve.
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
Effect of disturbance in Bia biosphere reserve 125
Acknowledgements Doucet, J.L., Kouadio, Y.L., Monticelli, D. & Lejeune, P. (2009)
Enrichment of logging gaps with moabi (Baillonella toxisperma
We are thankful to the Ghana-MAB committee, under the Pierre) in a Central African rain forest. For. Ecol. Manage. 258,
auspices of the Ghana Environmental Protection Agency 2407–2415.
(EPA) for permission to use inventory data collected as part Edwards, D.P. & Lawrance, W.F. (2013) Biodiversity despite
of the ‘Green Economy Project’ for the analysis. We are selective logging. Science 8, 636–647.
FAO (2010) Global Forest Resources Assessment. Food and
also grateful to the Manager and local mangement team of
Agriculture Organization of the United Nations, Rome.
the Bia biosphere reserve for assistance in Table S1
Forestry Commission (2010). Bia conservation area management
collection. plan. Wildlife Division, Accra, Ghana, pp 173.
Gardner, T.A., Barlow, J., Chazdon, R., Ewers, R.M., Harvey, C.A.,
Peres, C.A. & Sodhi, N.S. (2009) Prospects for tropical forest
References
biodiversity in a human-modified world. Ecol. Lett. 12, 561–582.
Addo-Fordjour, P., Obeng, S., Anning, A.K. & Addo, M.G. (2009) Gardner, T.A., Barlow, J., Sodhi, N.S. & Peres, C.A. (2010) A
Floristic composition, structure and natural regeneration in a multi-region assessment of tropical forest biodiversity in a
moist semi-deciduous forest following anthropogenic human-modified world. Biol. Conserv. 143, 2293–2300.
disturbances and invasion. Int. J. Biodivers. Conserv. 1, 21–37. Garrity, D.P. (2004) Agroforestry and the achievement of
Antwi, E.K., Boakye-Dankwa, J., Asabere, S.B., Yiran, G.A.B., the millennium development goals. Agrofor. Syst. 61, 5–17.
Loh, S.K., Awere, K.G., Abagale, F.K., Asubonteng, K.O., German Commission for UNESCO (2015). Management Manual for
Attua, E.M. & Owusu, A.B. (2014) Land use and landscape UNESCO Biosphere Reserves in Africa. A Practical Guide for
structural changes in the ecoregions of Ghana. J. Disaster Res. Managers, Bonn, Germany. ISBN: 978-3-940785-73-2
9, 452–464. Gotelli, N.J. & Chao, A. (2013) Measuring and estimating species
Appiah, M. (2011) Changes in plant species composition within a richness, species diversity, and biotic similarity from sampling
planted forest in a deciduous agroecosystem in Ghana. Agrofor. data. Encyclopedia of Biodiversity, 2nd edn. Vol. 5 (Ed S. A. LEVIN).
Syst. 85, 57–74. Academic Press, Waltham, MA.
Appiah, M. (2013) Tree population inventory, diversity and Hall, J.B. & Swaine, M.D. (1981) Distribution and Ecology of
degradation analysis of a tropical dry deciduous forest in Afram Vascular Plants in a Tropical Rainforest. Forest vegetation in Ghana.
Plains, Ghana. For. Ecol. Manage. 295, 145–154. W. Junk Publishers, The Hague, Netherlands.
Ardayfio-Schandorf, E., Yankson, P.W.K., Asiedu, A.B., Agyei- Hall, J.S., Harris, D.J., Medjibe, V. & Ashton, P.M.S. (2003) The
Mensah, S. & Attua, E.M. (2007). Socio-Economic Perspectives of effects of selective logging on forest structure and tree species
Off-Reserve Management in the Goaso Forest District of Ghana. composition in a Central African forest: implications for
Woeli Publishing Services, Accra. management of conservation areas. For. Ecol. Manage. 183,
Batisse, M. (1993) The silver jubilee of MAB and its revival. 249–264.
Environ. Conserv. 20, 107–122. Hall, A.R., Miller, A.D., Leggett, H.C., Roxburg, S.H., Buckling, A.
Bongers, F., Poorter, L., Hawthorne, W.D. & Sheil, D. (2009) The & Shea, K. (2012) Diversity– disturbance relationships:
intermediate disturbance hypothesis applies to tropical forests, frequency and intensity interact. Biol. Let. 8, 768–771.
but disturbance contributes little to tree diversity. Ecol. Lett. 12, Hammer, O., Harper, D.A.T. & Yan, P.D.R. (2001) PAST:
798–805. Paleontological Statistic software package for education and
Brose, U., Martinez, N.D. & Williams, R.J. (2003) Estimating data analysis. Paleontol. Eletron. 4, 1–9.
species richness: sensitivity to sample coverage and insensitivity Hansen, A.J. & Defries, R. (2007) Ecological mechanisms for linking
to spatial patterns. Ecology 84, 2364–2377. protected areas to surrounding lands. Ecol. Appl. 17, 974–988.
Burslem, D.F.R.P. & Whitmore, T.C. (1999) Species diversity Harvey, C.A., Komar, O., Chazdon, R., Ferguson, B.G., Finegan, B. &
susceptibility to disturbance and tree population dynamics in Griffith, D.M. (2008) Integrating agricultural landscapes with
tropical rainforest. J. Veg. Sci. 10, 767–776. biodiversity conservation in the Mesoamerican hotspot. Conserv.
Connell, J.H. (1978) Diversity in tropical rainforests and coral Biol. 22, 8–15.
reefs. Science 199, 1302–1310. Hawthorne, W.D. (1990). Field Guide to Forest Trees of Ghana.
Condit, R., Sukumar, R., Hubbel, S.P. & Foster, R.B. (1998) Natural Resources Institute for the Overseas Development
Predicting population trends from size distributions: a direct test Administration, Chatham, London Ghana Forestry Series 1.
in a tropical tree community. Am. Nat. 152, 495–509. Hawthorne, W.D. (1993). Forest Regeneration after Logging:
Defries, R., Hansen, A.J., Newton, A.C., Hansen, M. & Townshend, Findings of a Study in the Bia South Game Reserve Ghana. ODA
J. (2005) Isolation of protected areas in tropical forests over the Forestry Series 3. Inventory Project, Kumasi.
last twenty years. Ecol. Appl. 15, 19–26. Hawthorne, W.D. & Abu-Juam, M. (1995) Forest Protection in
Dickson, K.B. & Benneh, G. (1998). A New Geography of Ghana, Ghana: With Particular Reference to Vegetation and Plant Species.
Longman Group U.K. Limited, Harlow. IUCN, Switzerland.
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
126 Emmanuel Morgan Attua et al.
Hawthorne, W.D., Agyeman, V.K., Abu-Juam, M. & Foli, E.G. Ola-Adams, B.A. (1987) Effects of logging on the residual stands of
(2001). Taking stock. DFID report, Accra. a lowland rainforest at Omo Forest Reserve, Nigeria. Malays.
Hutchison, J., Dalziel, J.M., Keay, R.W.J. & Hepper, F.N. (1957- For. 50, 403–413.
1972). Flora of West Tropical Africa. 2nd edition. Vol. I-II, The Pappoe, A.N.M., Armah, F.A., Quaye, E.C., Kwakye, P.K. & Buxton,
Whitefriars Press, London. G.N.T. (2010) Composition and stand structure of a tropicalmoist
International Plant Nomenclature Index, IPNI (2008). Published semi-deciduous forest in Ghana. Int. Res. J. Plant Sci. 1, 095–106.
on the internet. Available at: http://www.ipni.org (Accessed on Ramirez-Marcial, N., Gonzalez-Espinosa, M. & Williams-Linera, G.
17 March 2014). (2001) Anthropogenic disturbance and tree diversity inMontane
Ishwaran, N., Persic, A. & Tri, N.H. (2008) Concept and practice: rainforests in Chiapas, Mexico. For. Ecol. Manage. 154, 311–326.
the case of UNESCO biosphere reserves. Int. J. Environ. Sustain. Sagar, R., Raghubanshi, A.S. & Singh, J.S. (2003) Tree species
Dev. 7, 118–131. composition, dispersion and diversity along a disturbance
IUCN (2010). A new vision for biodiversity conservation. Strategic gradient in a dry tropical forest region of India. For. Ecol.
plan for the convention on biological diversity (BCD) 2011– Manage. 186, 61–71.
2020. IUCN, Nogoya, Japan. Sahu, P.K., Sagar, R. & Singh, J.S. (2008) Tropical forest structure
IUCN & UNEP-WCMC (2014). The World Database on Protected and diversity in relation to altitude and disturbance in a
Areas (WDPA). UNEP-WCMC, Cambridge, U.K. Biosphere Reserve in central India. Appl. Veg. Sci. 11, 461–470.
Kent, M. & Coker, P. (2002) Vegetation Description and Analysis: A Schroth, G., Da Fonseca, G.A.B., Harvey, C.A., Claude, A., Gascon,
Practical Approach. Wiley, New York. C., Vasconcelos, H.L. & IZAC, A.-M.N. Eds. (2004). Introduction:
Kumar, R. & Shahabuddin, G. (2005) Effects of biomass extraction role of agroforestry in biodiversity conservation in tropical
on vegetation structure, diversity and composition of forests in landscapes. In: Agroforestry and Biodiversity Conservation in
Sariska Tiger Reserve, India. Environ. Conserv. 32, 1–12. Tropical Landscapes. Island Press, London.
Lamprecht, H. (1989). Silviculture in the Tropics, Tropical Forest Sheil, D. (1999) Tropical forest diversity, environmental change
Ecosystems and their Tree Species, Possibilities and Methods for their and species augmentation: after intermediate disturbance
Long Term Utilization. T2Verlagsgesells Chaft mbH, Postatch hypothesis. J. Veg. Sci. 10, 851–860.
1164, D6101 RoBdort, Federal Republic of Germany. Struhsaker, T.T., Struhsaker, P.J. & Siex, K.S. (2005) Conserving
Lauma-Aho, T. (2003). Natural Regeneration of African Mahogany Africa’s rain forests: problems in protected areas and possible
(Khaya ivorensis) in the Moist Semi-deciduous Forest in Ghana. solutions. Biol. Cons. 123, 45–54.
M.Sc. Thesis, The University of Helsinki, p. 63. Swaine, M.D. & Hall, J.B. (1988) The mosaic theory of forest
Makana, J. & Thomas, S.C. (2006) Impacts of selective logging and regeneration and the determination of forest composition in
agriculture clearing on forest structure, floristic composition and Ghana. Journal of Tropical Ecology 4, 253–269.
diversity, and timber tree regeneration in the Ituri Forest, Taylor, C.J. (1952). The Vegetation Zones of the Gold Coast, Forestry
Democratic Republic ofCongo.Biodivers. Conserv.15, 1375–1397. Department Bulletin No. 4, Government Printer, Accra.
Makundi, W.R. & Sathaye, J.A. (2004) GHG Mitigation potential Tilman, D. (1982) Resource Competition and Community Structure.
and cost in tropical forests - relative role of agroforestry. Environ. Princeton University Press, Princeton.
Dev. Sustain. 6, 235–260. Townsend, C.R. (1989) The patch dynamics concept of stream
Malhi, Y., Adu-Bredu, S., Asare, R.A., Lewis, S.L. & Mayaux, P. community ecology. J. N. Am. Benthol. Soc. 8, 36–50.
(2013) The past, present and future of Africa’s rainforests. Phil. Townsend, C.R. & Hildrew, A.G. (1994) Species traits in relation to
Trans. R. Soc. Lond. B. Biol. Sci. 368, 20120293. a habitat templet for river systems. Freshw. Biol. 31, 265–275.
McCullough, J., Decher, J. & Kpelle, D.G. (2005). A Biological UICN/PACO (2010). Parks and Reserves of Ghana: Management
Assessment of the Terrestrial Ecosystems of the Draw River, Boi- Effectiveness Assessment of Protected Areas. Burkina Faso:
Tano, Tano Nimiri and Krokosua Hills Forest reserves, Southwestern Ouagadougou.
Ghana. RAP Bulletin of Biological Assessment 36, Conservation Verburg, R. & Van Eijk-Bos, C. (2003) Effects of selective logging
International, Washington DC. on tree diversity, composition and plant functional type patterns
MEA (2005). Millennium Ecosystem Assessment. Ecosystems and in a Bornean rain forest. J. Veg. Sci. 14, 99–110.
Human Well-being: Biodiversity Synthesis. World Resources van der Werff, H. & Consiglio, T. (2004) Distribution and
Institute, Washington, DC. conservation significance of endemic species of flowering plants
Ministry of Environment and Science (2002). National Biodiversity in Peru. Biodivers. Conserv. 13, 1699–1713.
Strategy for Ghana. Ministry of Environment and Science, Whitmore, T.C. & Burslem, D.F.R.P. (1998). Major disturbances in
Republic of Ghana, Accra. tropical rainforests. In:Dynamics of Tropical Communities (Eds D.M.
Ministry, of Lands, and Forestry (1994) Forest and Wildlife Policy. NEWBERY,H.H.T. PRINS andN.D.BROWN).Blackwell Science,Oxford.
Republic of Ghana, Accra. Wiafe, E.D. (2014) Tree regeneration after logging in rain-forest
Myers, N., Mittermeier, R.A., Mittermeier, C.G., da Fonseca, G.A.B. ecosystem. Res. J. Biol. 2, 18–28.
& Kent, J. (2000) Biodiversity hotspots for conservation Wilkinson, D.M. (1999) The disturbing history of intermediate
priorities. Nature 403, 853–858. disturbance. Oikos 84, 145–147.
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127
Effect of disturbance in Bia biosphere reserve 127
Wittemyer, G., Elsen, P., Bean, W.T., Burton, A.C.O. & Brashares, Supporting information
J.S. (2008) Accelerated human population growth at protected
area edges. Science 321, 123–126. Additional Supporting Information may be found in the
Zhu, J.J., Mao, Z.H., Hu, L. & Zhang, J.X. (2007) Plant diversity of online version of this article:
secondary forests in response to anthropogenic disturbance
levels in montane regions of northeastern China. J. For. Res. 12, Table S1. Tree species, their families and abundance at
403–416.
three levels of anthropogenic disturbance in the Bia
biosphere reserve of southwestern Ghana.
(Manuscript accepted 27 March 2017)
doi: 10.1111/aje.12427
© 2017 John Wiley & Sons Ltd, Afr. J. Ecol., 56, 116–127