Assessment of floral composition, structure and natural regeneration of the Tano Offin Globally Significant Biodiversity Area of Ghana

Threats of forest degradation of the Tano Offin Globally Significant Biodiversity Area present the need to generate eco-information pertinent for its conservational purposes. Ten 50 m × 50 m plots (tree layer) were assessed for plant life forms with diameter ≥ 10 cm. A 10 m × 10 m plot (shrub layer) was located within each of the 50 m × 50 m plots where plant life forms with diameter < 10 cm were assessed. 1 m × 1 m quadrats (herb layer) were laid at the corners of the 50 m × 50 m plots and its centre for canopy closure and natural regeneration assessments. Plant species (240) belonging to 59 families were identified: 171 trees, 41 lianas, 11 shrubs, 7 herbs, 7 herbaceous climbers, 1 epiphyte, 1 grass and 1 fern. Species diversity (H´) of the tree, shrub and herb layers was 2.55, 2.54 and 2.48 respectively. The average maximum tree height was 46.19 m and the basal area was 28.36 m2/ha, which is below the 35 m²/ha conventional basal area value of tropical forests. Celtis mildbraedii and Rinorea welwitschii were the most structurally significant species at the tree and shrub layers, respectively, and a total of 75 tree species were regenerating.


Introduction
Tropical forest provides ecological resources (Antwi 1999) which translates into economic and sociocultural benefits (Abeney 1999). However, the impact of human activities has resulted in remarkable changes in species composition, abundance and diversity of organisms in various ecosystems including tropical forests (FOSA 2002, FRA 2010, Kim & Byrne 2006, Philip 1997. Reports by the 2010 Global Forests Resources Assessment indicated a 2 % annual forest loss from 1990 to 2000 in Ghana (FRA 2010). According to Global Forest Watch, the rate of deforestation exacerbated by 60 % in 2018 compared to 2017 with much of the destruction occurring within protected areas (Gbadamosi 2020). Deforestation of tropical forests threatens the sustainability of its biodiversity, demanding the conservation of remnant species lest these may be subjected to extinction (Myers et al. 2000). In this regard, floristic assessment of the forest is indispensable in detecting the risk of extinction, arrival of invasive species and changes in plant diversity over time.
Relatively, limited work has been done in determining the floristic composition and structure of forests in Ghana (Addo-Fordjour et al. 2009b, Anning et al. 2008, Hall & Swaine 1981, Hawthorne 1993, Pappoe et al. 2010, Vordzogbe et al. 2005. In 1981, an extensive assessment of the distribution of vascular plants in Ghanaian forests was carried out by Hall and Swaine which included the Tano Offin Forest Reserve (Hall & Swaine 1981). It has been decades now and there is thus the need to generate a carefully compiled up-to-date data on the floral composition of the reserve. More so, inventories by the Forestry Commission of Ghana in the reserve have focused mainly on timber species (Affum-Baffoe pers.comm) which is deficient for conservation purposes.
Furthermore, the Tano Offin forest has been a reserve of bauxite deposits (estimated at 700 million metric tons) which presents a potential threat to its forest diversity (Yoda 2020).
In the year 2019, the government of Ghana signed a memorandum with China to explore Ghana's deposits of bauxite found in Tano Offin and Atewa Forest Reserves, which constitutes two exceptional Upland Evergreen forests of Ghana (Gbadamosi 2020). These ecologically sensitive Forest Reserves are viewed as 'scientific gold mine' (Oteng-Yeboah, 2019), and research studies are critically needed to consolidate conservation demands.
With plant species composition and stand structure serving as important indicators in the formulation of conservation measures (Lindenmayer et al. 2000, Newton et al. 2003, information on the current state of the Tano Offin Globally Significant Biodiversity Area (GSBA) should be useful in enhancing conservation efforts of the GSBA. This is critical for maintaining healthy plant diversity and increasing resilience to environmental challenges (Shah 2009). The study focuses on the specific objectives of determining the floristic composition, structure, composition of natural regeneration and canopy closure of the Tano Offin GSBA. It is anticipated that the research findings will inform on the management and conservation policies of the forest.

Study site
The study was conducted in the Tano Offin GSBA which is a part of the Tano Offin Forest Reserve of Ghana. Tano Offin Forest Reserve falls within the Moist Semi-deciduous forest zone of Ghana with 34,100 ha of the reserve occurring as an Upland Evergreen forest (Birdlife International 2011, Ntiamoa-Baidu et al. 2001. A fraction of the Tano Offin Reserve was designated as a GSBA in 1999 after the discovery of the area's outstanding biological diversity of global conservation importance (FC 2007, McCullough et al. 2007). Using the Genetic Heat Index (GHI)an index of the concentration of rare plants within forest community, the GSBA concept was instituted to promote preservation of forests so as to protect unique flora, fauna and ecosystems (Asamoah et al. 2011, McCullough et al. 2007). The labelling as GSBA corresponds to IUCN's Category IV designation, that is, a protected area designated mainly for conservation through management intervention (McCullough et al. 2007).
The reserve lies between longitudes 1°57' and 2°17' West and latitudes 6°54' and 6°35' North, covering an area of 413.92km² of which the GSBA constitutes 44.5 %, that is, 178.34 km² (FC 2007). As an Important Bird Area, nationally rare bird species have been identified in this reserve and these include Cercococcyx olivinus Sassi, Columba unicinata Cassin and Tockus camurus Cassin (Birdlife International 2011, Ntiamoa-Baidu et al. 2001. The area experiences a bi-modal rainfall pattern where it peaks in May-June and September-October; the mean annual rainfall is 1250 mm and the annual mean relative humidity is 80 % (FC 2007).

Plot establishment
Using a compass and a measuring tape, 10 50 m x 50 m sample plots (termed tree layer) were randomly demarcated for the identification and measurement of all trees and lianas with dbh (diameter at breast height) ≥ 10 cm as well as the identification of other plant life forms such as herbaceous climbers. To ensure permanency of the plots, reflective poles were pegged at the corners of the plots, and all trees and lianas with dbh ≥ 10 cm were tagged and painted at the dbh level. A 10 m × 10 m plot (termed shrub layer) was located randomly within each of the 50 m × 50 m sample plots for the assessment of saplings (≤ 3 m high and dbh < 10 cm) and other plant life forms of dbh < 10 m. In addition, five single 1 m × 1 m quadrats (termed herb layer) were laid at the corners of the 50 m × 50 m plots and its centre for the assessment of forest floor vegetation and canopy closure. A Garmin GPS 76 was used to determine the geo-reference positions of sample plots (Supplementary Table 1).

Plant identification methods
Identification while in the forest was made possible by diagnostic factors (Supplementary Figure 1) such as growth habit, a study of the crown shape, tree bole structure, the bark texture and its slash appearance, the smell, taste and a study of the nature of exudates from the slashed bark, the leaves, fruits, and flowers. The aid of a catapult was employed to fetch tree leaves that were not easily within reach. The identity of plant species that could not be determined in the forest were identified with the aid of Hawthorne & Jongkind (2006), Hawthorne & Gyakari (2006) and the herbarium of the Resource Management Support Centre of the Forestry Commission of Ghana.

Tree diameter and height measurement
Within the 50 m × 50 m plots, the dbh of all trees and lianas with dbh ≥ 10 cm were measured using a tree caliper. The dbh of trees with buttresses were measured with a relascope. The diameter of saplings (≤ 3 m high and dbh < 10 cm) and other plant life forms of dbh < 10 cm were also measured in the 10 m × 10 m plot. Within the 50 m × 50 m, the height of all trees and shrubs with dbh ≥ 10 cm was determined with Vertex IV and Transponder III. Trees with broken tops were noted and dead trees were eliminated.

Canopy measurement and regeneration assessment
The concave spherical densiometer (Model C) was used to assess the forest canopy closure. Readings were taken on the 1 m × 1 m quadrats located at the corners and centre of the 50 m × 50 m sample plots. Based on recommendations from literature (Fiala et al. 2006, Jennings et al. 1999, readings were taken at each of the cardinal direction, resulting in four readings for each point of measurement or quadrat. Seedlings (< 1.5 m high and dbh ≤ 1.5 cm), grasses, herbs and other forest floor vegetation found within the 1 m × 1 m quadrats were identified and counted.

Data analysis
Species richness for each plot was determined, and the total number of genus, families and life forms were assessed. To describe the numerical structure of the forest, diversity quantification tools Simpson's (1-D) and Shannon-Wiener diversity [H´= −∑(Pi*lnPi)] indices were used, where ∑ is summation, Pi is the proportion of individuals that belongs to species i, and ln is the natural log; D = ∑(n/N) 2 with n being the number of individuals of each species, and N is the total number of individuals of all species. The basal area (G = πr 2 ) per hectare for both plant species with dbh ≥ 10 cm (tree layer) and those with dbh < 10 cm (shrub layer) was computed. The structural significance of tree species was assessed by calculating the Importance Value Index (IVI) of each species as shown below: Importance Value Index ¼ðrelative density þ relative dominance þ relative frequencyÞ The ecological guild (pioneer, non-pioneer light-demanding and shade-bearing) and star rating of the tree species were determined using Hawthorne (1993) and Hawthorne & Gyakari (2006). The star rating consists of black star species (rare internationally and at least uncommon in Ghana); gold star (fairly rare internationally and locally); blue star (widespread internationally but rare in Ghana or vice-versa); scarlet star (common, but under serious pressure from heavy exploitation); red star (common, but under pressure from exploitation); pink star (common and moderately exploited as well as being non-abundant and of high potential value); and green star species (common in Ghana and of no particular conservation concern). In accordance with Antwi (1999), trees of dbh ≥ 10 cm were classified into four height classes (emergent: > 35 m, upper canopy: > 25-35 m, lower canopy: >15 m-25 m and understorey: ≤ 15 m) to reflect the forest's vertical layer and the species composition of the various layers. The percentage canopy closure for each plot (i.e. 50 m × 50 m) as described by Marchi & Paletto (2010) was calculated. Abundance of seedlings (young tree plant with height < 1.5 m and dbh ≤ 1.5 cm) and saplings (young tree plant with height ≤ 3 m and dbh < 10 cm) were also determined.

Floristic composition
A total of 240 plant species were identified during the assessment (Supplementary Table 2). These comprised of 171 trees, 41 lianas, 11 shrubs, 7 herbs, 7 herbaceous climbers and a species of an epiphyte (Microsorum punctatum), grass (Leptaspis cochleata) and a fern (Adiantum vogelii). The star rating of 222 plant species were determined: 1 black star, 8 gold stars, 17 blue stars, 9 scarlet stars, 7 red stars, 21 pink stars and 159 green stars; the star ratings of 18 plant species were not available (Supplementary Table 2). With the exception of 33 plant species, the guilds of 207 species were determined, namely 42 pioneers, 68 non-pioneer lightdemanding species, 93 shade-bearing, 2 swamp species and 2 invasive species (Supplementary Table 2). There were 179 genera and 59 families with Fabaceae being the most contributing family with respect to species richness (15 %)represented by 36 species (Figure 1). Forty-six families had less than five species (Figure 1). Out of the total 240 plant species, 97 occurred just once (i.e. among the 10 main plots) while Celtis mildbraedii, Culcasia angolensis and Strombosia pustulata occurred on all the 10 main plots ( Figure 2). Incidence of plants species decreased with increase in the number of plots such that fewer species were common on most plots ( Figure 2).

Forest structure
Tree layer Measurement in the tree layer constituted lianas and trees with dbh ≥ 10 cm and height > 3 m which were sampled from the 50 m × 50 m main plots. A general summary of the findings of the tree layer is in Table 1.
Diameter class distribution and basal area. The number of individuals in the diameter classes decreased with increasing diameter so that the highest number of plants with dbh ≥ 10 cm was found in the 10-30 cm diameter class (Figure 3). Species found in the > 90-110 cm class include Antiaris toxicaria, Celtis mildbraedii, Parkia bicolor, Petersianthus macrocarpus and Sterculia oblonga. Species of the highest diameter class (> 110 cm) include Alstonia boonei, Hexalobus crispiflorus, Parkia bicolor, Sterculia oblonga and Triplochiton scleroxylon.
Height. With respect to classifying individual trees of dbh ≥ 10 cm into height classes, four classes were obtained, namely the   Structural significance of plant species. With the highest density, maximum percent score of frequency (most common) and the greatest dominance, Celtis mildbraedii was the most significant species for the tree layer of the Tano Offin GSBA, recording an IVI of 32.16 (Supplementary Table 3). This was followed at a wider margin by Strombosia pustulata and Hymenostegia afzelii which recorded an IVI of 12.01 and 11.31, respectively. Out of the total 154 species, 53 were available on just a single occasion (i.e. density = 1; frequency = 2.5 %) so that the variation in IVI among these species resulted from differences in only their dominance. Dalbergia saxatilis and Monodora myristica were the least significant species occurring in the study area (Supplementary Table 3).

Shrub and herb layers
The shrub layer constitutes plant life forms with dbh < 10 cm and height ≤ 3 m that were sampled from the 10 m × 10 m plots. The herb layer refers to the forest floor vegetation which had the height of < 1 m and dbh of ≤ 1.5 cm, captured from the 1 m × 1 m quadrats. A general summary of the findings of these layers is in Table 1.
Diameter class distribution and basal area. In a similar pattern to the tree layer, there was a decrease in the number of individuals in the various diameter groups as sizes increase so that the highest number of plants with dbh < 10 cm was found in the smallest diameter class ( Figure 5). The seven species that formed the largest diameter class in the shrub layer (> 8-9.9 cm) are Blighia sapida, Celtis mildbraedii, Cola boxiana, Dacryodes klaineana, Napoleonaea vogelii, Rinorea oblongifolia and Rinorea welwitschii.
Structural significance of plant species. With an IVI of 33.94, Rinorea welwitschii was the most significant species for the shrub layer in the GSBA, being the most dominant and most abundant; it was followed at a wider margin by Drypetes chevalieri (13.46), Strombosia pustulata (11.46) and Greenwayodendron oliveri (10.37) (Supplementary Table 4). However, Rinorea oblongifolia with an IVI of 8.57 was the most frequent (70 %). With a density of 20 individuals per 0.1 ha and 10 % frequency level, Cleidion gabonicum appeared locally abundant (per plot) unlike Diospyros ferrea which was relatively widespread (40 %) though of lower abundance (5 individuals per 0.1 ha). Although Treculia africana occurred just once, it was far more significant than 75 other species due to its higher dominance. Pterygota macrocarpa, Rothmannia whitfieldii, Coffea stenophylla and Trilepisium madagascariense had the lowest IVI stemming from their small dominances (Supplementary Table 4).

Natural regeneration and canopy closure
A total of 75 plant species were found regenerating as saplings and seedlings. Species richness of saplings was 61, while 45 species were seedlings (Supplementary   However, the precise growth form classification for some plants was challenging, since some plant species change their habit in accordance with growing conditions. Uvaria and Combretum species, for instance, behave and flower as shrubs when growing in open vegetation but become lianas when they live long enough, and forests develop around them (Hawthorne & Jongkind 2006). In this study, there was identification disparity for some species such as Dichapetalum angolense which was found growing sometimes as liana and sometimes as a shrub or tree. The preponderance of Fabaceae family is similar to findings in other tropical forests: a Moist Semi-deciduous forest of Ghana  (Martinez & Phillips 2000) which is similar to 59 families/2.5 ha in this study, with Fabaceae representing~19% akin to the 15% representation in this study. Apart from Australia, the Fabaceae family together with the following families found in this study are also reported to be commonly found in all tropical forests: Rubiaceae, Annonaceae, Euphorbiaceae, Moraceae, Sapotaceae, Myristicaceae and Meliaceae (Banin et al. 2015).
The presence of seven red stars and nine scarlet stars could be an indication of serious exploitation activities in the GSBA. Pink stars which are moderately exploited were slightly higher in number (21) than the scarlet (9) and red stars (7). Their higher presence than red and scarlet stars could also be that the values of some pink stars are yet to be ascertained (Hawthorne et al. 1997). The incidence of a smaller number of plant species with respect to increase in plot number (Figure 2) is consistent with the observation of Magurran (1988) who attested that majority of plant species are rare within a normal ecological community, a moderate number are common, while only few are very common. While 97 species occurred just once (plot size of 0.25 ha) in this study, Morandi et al. (2016) found 142 species out of a total 257 species sampled from 10 plots (1 ha each) to be restricted to only one plot in the Cerrado and Amazonian domains in Brazil.

Forest structure
On per hectare basis, the herb layer contained the highest number of individual and species, followed by the shrub layer which was also higher in number and species than the tree layer. Similarly, there were more individual plant species at the shrub layer than they were in the tree layer at the Campo-Ma'an rain forest of Cameroon (Tchouto et al. 2004). The greater number of individuals and species in the shrub layer as compared to the tree layer may be due to the many growth forms which make up the shrub layer, namely shrubs, shrublets, small trees (pigmy trees and treelets), immature large trees, liana and hemi-epiphytes, of which some are not usually found in the tree layer.

Species diversity
Diversity indices give a quantitative view of diversity and thus provide information about rarity and commonness of species in a community which is essential for understanding community's numerical structure (Beals et al. 1999). Although Shannon-Wiener and Simpson's diversity indices assess different facets of diversity through the relative weighting assigned to evenness and species richness (Magurran 1988), these diversity indices indicated that floristic diversity in all three layers were comparable. With regard to the Simpson's index of diversity values, the tree layer may be more even in its diversity outlook compared with the other two layers. Khaine et al. (2017) found the Shannon and Simpson's diversity indices of a tropical deciduous forest in Myanmar to increase with increasing rainfall. In another tropical forests, a Shannon diversity index range of H´= 2.74-2.99 is recorded to be indicative of an intermediate type of secondary succession, while a range of H´= 3.37-3.86 represented an advanced type of secondary suggestion (Jiménez et al. 2016). However, the Shannon diversity index value reported in this study is similar to the Shannon diversity value (H´= 2.9) reported for a degraded Moist Semi-deciduous forest in Ghana, where a higher index value (H´= 3.6) was reported for its undegraded portions, suggesting degradation, other than a level of secondary succession stage of the Tano Offin GSBA.

Diameter class distribution
Results of other studies conducted in some Moist Semi-deciduous forests of Ghana (Addo-Fordjour et al. 2009b, Pappoe et al. 2010 agree with findings in the Tano Offin GSBA where there was a similar decrease in the number of individuals in the various diameter groups as tree and liana diameter sizes increased so that the highest number of trees and lianas with dbh ≥ 10 cm were found in the 10-30 cm diameter class. This observation depicts the typical reverse J-shaped diameter class distribution curve of natural tropical uneven-aged forests. In addition, since, tree diameters could suggest tree age (Andreu et al. 2009), the abundance of trees and other plant forms in the lower diameter class size imply a dynamic self-regenerating stand. The disparity in basal area of tropical forests is attributable to differences in species composition, tree age, elevation, stage of succession and the level of anthropogenic disturbances (Gopalakrishna et al. 2015, Jiménez et al. 2016, Kacholi 2014, Khaine et al. 2017, Naidu & Kumar 2016. Earlier records of the mean basal area for the Upland and Moist Semi-deciduous forests of Ghana were 26.8 m 2 /ha and 24.2 m 2 /ha, respectively, which was attributable to high incidence of farming disturbance and logging (Hall & Swaine 1981). With reference to the 35 m²/ha average value for basal area of tropical forests (Philip 1983), the recorded 28.36 m²/ha basal area of the Tano Offin GSBA in this study can be described as below average. Over the years, there has been a decline in basal area as a result of logging of the Tano Offin Forest Reserve even before the GSBA was demarcated: between the years of 1990 and 2000, 5 % or 26.76 ha of forest cover of Tano Offin were lost, amounting to 2.76 % reduction in basal area (Djabletey 2005). The annual forest loss of Tano Offin Reserve in general is pegged at 0.3 % and as a result, basal area declined from 24.3 m²/ha to 18.9 m²/ha between the years of 1990 and 1996 and then it further declined from 18.9 m²/ha to 16.9 m²/ha between 1996 and 2001 (Djabletey 2005). Establishment of the GSBA with logging prohibition may have appreciated the basal area to some extent, although the prohibition it is not completely adhered to (Afriyie 2010, Asamoah et al. 2011, The Chronicle 2011. Similarly, a structural analysis of a Moist Evergreen tropical forest of Cameroun presented a basal area of 32 m²/ha and 29.4 m²/ha for the part of the forest where past logging activities have happened once and twice, respectively, while the unlogged section recorded a higher basal area of 35.3 m²/ha (Njepang 2015).

Height
Average tree height of the emergent layer was 46.19 m, similar to the report by FC (2007) where the average maximum height of trees in the reserve was found to be about 45 m. Past findings on the maximum tree height of the Moist Semi-deciduous forests by Hall & Swaine (1981) was > 50 m, while the tallest trees of the Upland Evergreen forests were found rarely to exceed 45 m. Greater constituents of the emergent in the GSBA were non-pioneer light-demanders which are known to include most of the timber species (Wong 1989). Pioneers were least represented in the emergent layer (> 35 m) as they hardly exceed the height of 20-30 m (FAO 2002) and they decline in numbers with respect to increasing size (Hall & Swaine 1988). Most constituents of the understorey were shade-bearers which flourish beneath upper and lower canopies. What could have also compounded the great number of shade-bearers at the understorey might be the presence of cryptic pioneers which are often misclassified as shade-bearers; unlike pioneers, they tolerate shade later in life (Hawthorne 1993). In a cross-continental comparison of forest structure from intact closed-canopy tropical forest, the mean tree height for trees with 10 cm mean stem diameter was 13.3 m, 11.9 m and 10. 6 m for Central Africa, Asia (Borneo) and the Amazonia (Central/ East), respectively; that of trees with 40 cm mean stem diameter was 30.8 m, 30.3 m and 26. 1 m, respectively; and that of trees with 100 cm mean stem diameter was 43.5 m, 46.0 m and 39. 0 m, respectively (Lewis et al. 2013).

Structural significance
The representation of Celtis mildbraedii as the most significant species in the tree layer is consistent with the reported feature of the Moist Semi-deciduous forests of Ghana: Celtis mildbraedii together with Triplochiton sclerexylon are known as the most common species of the Semi-deciduous forests of Ghana, although the latter is less frequent (Hall & Swaine 1981). Pterygota macrocarpa, Mansonia altissima and Terminalia superba are reported to be the dominant timber species of the GSBA (FC 2007). Hence, the less favourability of C. mildbraedii for timber, compared to P. macrocarpa, M. altissima, T. superba and T. sclerexylon perharps must have appreciated its significance. In conformity to exploitation pressure, M. altissima, T. superba, P. macrocarpa and T. sclerexylon have all been classified under the reddish star system (Hawthorne 1993). Their exploitation must have necessitated the creation of P. macrocarpa, M. altissima and T. superba plantations in a Taungya system at the reserve (Birdlife International 2011). The significance of Celtis mildraedii was followed at a wider margin by Strombosia pustulata, and then Hymenostegia afzelii. When present, Strombosia pustulata is reported to have high incidence (Hawthorne & Jongkind 2006) and thus not surprising that S. pustulata recorded the highest frequency value after Celtis mildbraedii. As a characteristic species of the Upland Evergreen forest, Hall & Swaine (1981) similarly recorded a 100% frequency on Strombosia pustulata. Hymenostegia afzelii when present is reported to occur in abundance (Hawthorne & Jongkind 2006). Therefore, it is also not surprising that it recorded the highest value for density after Celtis mildraedii. Dalbergia saxatilis and Monodora myristica were the least significant species because they were the least dominant species.
Rinorea welwitschii was the most significant species for the shrub layer. Members of the Rinorea family are noted to be locally dominant and widespread understorey species (Hawthorne & Journal of Tropical Ecology 7 Jongkind 2006). This assertion is further confirmed in this study by the highest frequency score attained by Rinorea oblongifolia. The local abundance of Cleidion gabonicum (all 20 individuals were found on a single plot) at the shrub layer concurs with the view of Hawthorne & Jongkind (2006) who found that Cleidion gabonicum is a gregarious understorey tree. Overall, IVI is useful for prioritising plant species conservation (Kacholi 2014).

Natural regeneration
Less than 50 % of the adult population were regenerating as saplings and seedlings in this study. However, findings of Hall & Swaine (1988) indicated 68 % of the adults in their study regenerating and concluded that the Ghanaian forests were generally well represented by juveniles. Saplings had higher species richness than seedlings and a higher value of species diversity index. Most forest trees have recalcitrant seeds and are very unlikely to be found in seed banks. Emergence of seedlings, therefore, might have been challenged by factors such as removal of parent trees and lack of effective dispersal mechanisms. For a successful natural regeneration, a seed source, a suitable microclimate, light, freedom from vegetation competition and browsing are very crucial (Hale 2004, Ward & Worthley 2004. Constituents of the regeneration flora which occurred exclusively as saplings and seedlings were comparable to findings of Addo-Fordjour et al. (2009b), where seven species of the regeneration vegetation were missing from the adult tree population at Tinte Bepo Forest Reserve (Moist Semi-deciduous forest). Seedlings of Lovoa trichilioides if present are often found within the vicinity of parent trees (Hawthorne & Jongkind 2006). The red star status of the species (Hawthorne 1993) suggests exploitation pressure on adult trees which might have led to their absence among the adult trees in this study. Lecaniodiscus cupaniodes and Mallotus oppositifolius may have been new to the community as colonisers. Although Lecaniodiscus cupaniodes is purported to be a cryptic pioneer (regenerating in gaps under canopy), both are classified as shade-tolerant species. They are reported to be very common in secondary forests with Mallotus oppositifolius exhibiting weedy behaviour (Hawthorne & Jongkind 2006).

Canopy closure
A forest canopy is considered open when 10-39 % of the sky is obstructed by tree canopies, moderately closed when 40-69 % of the sky is obstructed by tree canopies and closed when 70-100 % of the sky is obstructed by tree canopies (O'Neil et al. 2001, Portland State University: PSU (2010. The forest canopy of the Tano Offin forest GSBA could thus be described as closed. The structure of the forest is a likely contributing factor to the closed forest canopy of the GSBA. The structure of forests determines to a large extent the amount of light transmitted to the forest floor. The basal area of a stand with many small trees will have a very dense canopy and will transmit less light (greater canopy closure) than a stand with the same basal area but fewer, larger trees (Hale 2004). The closed canopies of the GSBA may have favoured the shade-bearers and non-pioneer light-demanding species that composed a greater portion of the natural regeneration flora.

Conclusion
The study has largely ascertained the threat that forest degradation poses to the Tano Offin GSBA. If serious attention is not given to manage the floral diversity of the GSBA, its conservation value will decrease. Therefore, the star rating system should be updated, especially the reddish star system. This is critical in reflecting the current exploitation pressure on the plant species, so as to accord the necessary conservation attention. Information from the forest structure analysis suggests abundance of young trees to rejuvenate the GSBA into a functionally matured forest in the years to come if degradation on the area is curbed. By the establishment of permanent sampling plots, the study serves as a baseline for future research that pertains to understanding the dynamics of the changing forest resources as well as monitoring impacts of anthropogenic disturbances. Overall, the information generated should be useful in designing conservation measures for the Tano Offin GSBA as well as helpful in managing protected areas.