Assessing plant diversity change in logged and unlogged dense semi-deciduous production forest of eastern Cameroon

Abstract

This study was carried out in the dense semi-deciduous production forest of East Cameroon. The objective of this work of this study was to provide comparative floristic knowledge that can serve as a basis for the planning and sustainable management of ligneous plant resources in Communal Forests before and after logging. Sampling was done in unlogged and logged forest. Data collection was carried out using the linear transects subdivided into 10 plots of 25 m × 20 m (500 m²) with an equidistance of 225 m for the inventory of all trees with dbh ≥10 cm installed measured at 1.3 m above ground level. Nested quadrats 5 m × 5 m, oriented south-west and north-east were set up in each plot for the counting and identification of all individuals with a diameter less than 10 cm. The analysis of inventory data showed that the floristic composition was higher in the unlogged forest. The individuals were more evenly distributed in the logged (Pielou's equitability index = 0.83) than in the unlogged forest. The study of the functional spectra showed that the flora of the two forest types was dominated by Guinean-Congolese species (67.57% in the unlogged forest and 63.07% in the logged forest) and Phanerophytes, particularly Mesophanerophytes. The dominance of sarcochorous species reveals that the main mode of dissemination in this forest is zoochory, particularly endozoochory. The presence of pleochroic species in the logged forest reveals the importance of dissemination by water in the environment. The plants surveyed were divided into five plant assemblages (three for logged forest and two for unlogged forest) based on ecological parameters. The findings of this study suggest that forest management which combines assisted natural regeneration with the natural process of secondary succession facilitates the reconstitution of the vegetation cover and, by extension, the conservation of biodiversity in post-logging forest concessions.

1. Introduction

The assessment of Cameroon's forest cover shows that it is 22.5 million hectares [1], occupying 48% of the national territory [2]. The forest typology based on land use classes highlights two major categories: dense rainforests and other forest types [3]. Dense rainforests are the most diverse and account for nearly 60% of total biodiversity [3,4]. They are divided into lowland dense evergreen rainforests (54% of the forest area) and lowland semi-deciduous rainforests (28% of the forest area) [5]. The latter, because of its rich diversity of commercially important tree species and relatively small area, is more threatened than the dense evergreen forest. These forests are important reserves of genetic, specific and ecosystem diversity that should be conserved as much as possible for the sustainable management of this biological heritage [[6], [7], [8]]. These forests provide multiple services [9,10], contribute to the regulation of the greenhouse effect and the establishment of major climatic balances [11,12]. Thus, these forest ecosystems represent useful conservation areas for humanity because of the role they play in maintaining biodiversity [13]. However, these forest ecosystems are subject to various disturbances generated mostly by human activities [[14], [15], [16]]. These anthropogenic disturbances through logging, harvesting of non-timber forest products (NTFP), slash and burn shifting cultivation, mining and infrastructural developments [17], lead to a decrease in the forest area [18] and the loss of biodiversity. All these threats to forests and biodiversity are a call for concern to the national and international communities which seek to develop strategies that reduce emissions from deforestation and degradation of forests.

However, the Cameroonian government, cognisant of the role of its biodiversity and the threats that lead to its degradation, has set up regulatory frameworks to perpetuate a sustainable production of its forestry resources. This desire is reflected in the adoption of the Forestry Law in 1994 [19], which governs forests, wildlife and fisheries. All these initiatives underpin the sustained production of timber, the conservation of biodiversity and the maintenance of forest functions. Thus, these reforms contain innovative elements such as the involvement of local populations in forest management, the zoning of the forest estate and the attribution of new forest exploitation titles to councils for ownership and management of Communal Forests (CF). Following the 1994 zoning, Cameroon's forests are divided into non-permanent and permanent forest estates, of which the CF are part. However, the CF, which are allocated for a thirty-year rotation and are intended for the sustainable management of natural resources, are of particular interest for decision-makers. Those under management in the Congo Basin are of increasing interest in the REDD + mechanism [20,21] as they may attract more benefits from the carbon market. Reconciling forest conservation and exploitation in order to preserve their ecological role and to satisfy the needs of local populations, is a real challenge for Communal Forests. Therefore, the sustainable management of communal forests which is limited in practical terms should be based on the knowledge and ecological functioning of its resources [22,23].

Currently, the problem in the sustainable management of Communal forests is the lack of comparative studies that provide knowledge gaps necessary to be addressed in communal forest management plans and more precise for local capacity building in a context of participatory management to ensure conservation and sustainable use of biodiversity. However, the study of floristic composition, species diversity, and structural analysis is crucial to provide the necessary information on species richness and diversity in forests, and vegetation types are useful for forest management and help understand forest ecology and ecosystem functions [23] before and after logging. Previous works focused mostly on species diversity, conservation and floristic composition [4,13,15,[24], [25], [26], [27]]. However, works dealing with comparisons of floristic composition and species richness between a logged and unlogged plot in dense semi-deciduous production forests are only rarely discussed.

Thus, current knowledge of these tropical forests lacks quantitative data, which are particularly crucial for estimating the beta diversity in the different plots (logged and unlogged) and for ensuring good management of the plant resources. The availability of information on the natural heritage and its evolution is a fundamental element for the implementation of the national strategy for the conservation and sustainable management of biodiversity. Studies have revealed that although tropical forests disappear rapidly through deforestation, they also have the potential to regrow naturally through the process of secondary succession [28]. Hence, the aim of this study is to provide comparative floristic knowledge that can serve as a basis for the planning and sustainable management of ligneous plant resources in Communal Forests before and after logging. More specifically, the objectives of the study were: to assess the impact of industrial logging on ligneous plant diversity, and to determine the importance of secondary succession in the reconstitution of the vegetation of dense semi-deciduous forest after logging activities. This study is based on the hypothesis that secondary succession favors the reconstitution of logged tropical forests affected by low intensity industrial logging that modifies the floristic and ecological characteristics of the forest.

2. Material and methods

Dimako Communal Forest is located in the East Region of Cameroon Region, in the Upper Nyong Division, Dimako District (Fig. 1). It is located between latitudes 4°10' and 4°20' North and longitudes 13°30' and 13°50' East and has a surface area is 16200 ha. The area is subject to the Guinean equatorial type of climate with an annual succession of four seasons. The average rainfall varies between 1.500 and 1.800 mm/year and the average temperature is 24 °C. The slopes range from 0 to 15% and the altitude varies between 596 and 689 m [29]. Phytogeographically, its vegetation is that of the dense semi-deciduous forest [30]. The forest forms part of the Guinean-Congolese dense semi-deciduous forest with Malvaceae and Ulmaceae being the characteristic families [31].

Fig. 1.

Map of study site.

2.1. Collection of floristic inventory data

The logged forest and unlogged forest all belong to the same forest massive and have the same fundamental ecological conditions. This is the basis for which the synchronic analysis was done.

Sampling was done in unlogged and logged forests of the Communal Forest using the same experimental design. The selection criterion for the unlogged forest was the absence of anthropogenic disturbance. A logged forest (15 years after the passage of industrial logging) was selected for this study. The logging intensity was low, approximately 0.78 harvestable stems per hectare [29]. After logging, further disturbance of the logged plot mainly due to anthropogenic action was prohibited. The study focused on woody plant species diversity because logging primarily targets woody plants species and causes damage to the surrounding vegetation which can impact on the diversity of these plants [32].

A representative and homogeneous vegetation was selected through an analysis of satellite images (Sentinel 2A, 2015) and topography. The sampling design was a linear transect of 2.5 km subdivided into 10 plots of 25 m × 20 m (500 m²) with an equidistance of 225 m for the inventory of all trees with dbh ≥10 cm. Two nested quadrats of 5 m × 5 m were installed in each plot, one southwest and the other northeast for the counting of all plant individuals 1 cm < dbh <10 cm. A total of 80 plots of 25 m × 20 m (4 ha) and 160 quadrats of 5 m × 5 m for the logged forest and the unlogged forest were sampled in this study, that is, 40 plots of 25 m × 20 m and 80 quadrats of 5 m × 5 m in each forest type. The transects were installed randomly following the vegetation types predefined using the analysis of satellite images.

For each tree counted, parameters such as: stem diameters, scientific or vernacular names and/or trade names were recorded. To limit the risk of double counting of woody individuals in forest plots, all trees were marked with stencil paint. The geographical coordinates of each plot were recorded using a GPS. Herbarium samples of the tree species unidentified in the field were collected, pressed and preserved in 70% alcohol. These were later confirmed and identified at the Cameroon National Herbarium in Yaoundé. The taxonomic nomenclature adopted in this study was the Angiosperm Phylogeny Group (APG IV).

2.2. Alpha diversity

2.2.1. Species richness

The floristic composition was assessed on the basis of species richness, genus and family diversity [33].

2.2.2. Shannon diversity index

This index was assessed using the mathematical formula:

H’ = $-\sum \frac{\mathrm{n}\mathrm{i}}{\mathrm{N}}\ln \frac{\mathrm{n}\mathrm{i}}{\mathrm{N}}$ [34],Where H' is the Shannon index; ni is the number of individuals of species i; N is the total number of individuals of all species.

2.2.3. Simpson's diversity index

The Simpson’s diversity index was computed using the mathematical formula:

D = 1-$\sum _{i=1}^s{p}_i^2$

.

With pi: the number of individuals of species i; S: the total number of individuals of all species.

2.2.4. Pielou’s equitability index

The Pielou’s equitability index was computed using the mathematical formula:

E = $\frac{\mathrm{H}’}{\mathrm{l}\mathrm{n}\mathrm{S}}$ [35], where E is the Pielou’s equitability index; S is the total number of species in a plot.

2.2.5. Fisher-alpha diversity index

Fisher’salpha diversity was calculated using the mathematical formula:

S = αln (1 + N/α), where S: species richness (number of taxa), N: number of individuals; α: Fisher’s alpha index.

2.3. Beta diversity

The beta diversity was used to complement the alpha diversity and to account for diversity on a regional scale. In this study, the Sorensen's similarity coefficient was used measure the Beta diversity using the following mathematical formula: Ks = $\frac{2\mathrm{C}}{\mathrm{A}+\mathrm{B}}$ x 100 [36], where Ks is Sørensen's coefficient of similarity; A = the number of species in a list belonging to a plot; B = the number of species in a list belonging to a plot; C = the total number of species common to both plots to be compared.

2.3.1. Importance value index and Family Value Index

The dominance of species and families were assessed using the Importance Value Index (IVI) and the Family Value Index (FIV). The calculation of these indices was based on the determination of the following parameters: relative frequency, relative dominance and relative density.

Relative frequency = (Number of plot containing X specie/Total number of plot) X 100.

Relative abundance = (Number of individuals of the specie/Total number of individuals) X 100.

Relative dominance = (Total basal area of the specie/Total basal area of all specie) X 100.

Basal area (BA) is the area in square meters (or square feet) of the cross-section of the trunk of a tree at breast height (1.3 m or 4.5 ft) and it is most commonly used as an indicator of stand density and is expressed as square meters per hectare or square feet per acre. The basal area of an individual tree is related to its volume, biomass and crown parameters. The basal area of individual trees in the sample plots were calculated using the formula:

BA= (πD²)/4.Where: BA = basal area (m²), D = diameter at breast height (cm) and π = pi (3.142).

The Importance Value Index was calculated using the formula of [37] as follows:

IVI = Relative frequency + Relative abundance + Relative dominance.

Where IVI = Importance Value Index.

The family Importance Value index was computed as the sum of the relative density, dominance and relative frequency of each family as follows:

FIV = [(Number of species of family X/Total number of species) x 100 + (Number of individuals of family X/Total number of individuals) x 100 + (Sum of basal area of family X/Total basal area) x 100)].

2.4. Generic coefficient (GC)

The generic coefficient was computed as follows:

GC = Number of genera

$$GC=\frac{Numberofgenera}{Numberofspesies}\times 100$$

2.5. Ecological spectra

In order to study the functional diversity of the forests, each forest type was characterised according to various ecological spectra that could provide information about a changing environment. This information allows for better understanding of the role of each species in the ecosystem and their proportion represents the response of the vegetation to changes that have taken place either on large scales of time and space (biogeographical characteristics) or on shorter scales (biological types responding to recent disturbances). The functional parameters of species in an ecosystem reflect their response to environmental disturbances [38].

2.6. Biological types

The examination of biological types allows for the determination of adaptive strategies as well as the physiognomy of the vegetation. The biological types defined according to the classification of [39] and modified by Ref. [40] were adopted for this study.

2.6.1. Types of diaspores and modes of dispersion

The dispersion data show the ability of plants to disperse in the forest and their potential to colonise new environments. Diaspora types and modes of spread were defined according to the classification of [41] (1957) and [42]. The modes of dissemination were identified using the morphological attributes developed by the species to disperse the seeds [41].

2.6.2. Species temperaments

According to Ref. [43], temperament refers to the overall growth and development characteristics of a tree in a forest eco-unit. For tree species, colonisable microhabitats are defined in relation to the dynamics of canopy renewal [44].

2.6.3. Phytogeographical distribution

The phytogeography of the species inventoried makes it possible to determine the geographical distribution of the species within the forest landscape. Central Africa has been divided into several parts that provide information on the phytogeographical distribution of species [45]. This chronological subdivision was completed by Ref. [46] for the Lower Guinean sub-centre.

Hierarchical ascending classification was carried out in order to identify the discriminating ecological factors and the floristic gradients linked to them. The means of the diversity indices were compared between plots, using one-way analysis of variance (ANOVA) tests. Where the differences between the means of the parameters were significant (p < 0.05), post-hoc tests (Tukey test and Kruskal-Wallis test) were carried out to identify which groups were different. To carry out the various statistical analyses, the Excel spreadsheet was used to produce the figures and tables; PAST 4.0 and R.4.1.2 softwares were used for the descriptive analyses.

3. Results

3.1. Ordination of the floristic records of the plots

The matrix of 80 records in the unlogged and logged forest subjected to Principal Component Analysis (PCA) discriminates 5 groups of records (Fig. 2). The records are distributed according to the relationships between species, records and groups formed. In the unlogged foorest, three plant groups were identified while in the logged forest, two plant groups were identified.

• −Group 1 consisted of 70 records in the unlogged forest. The characteristic species were Cordia platythyrsa Baker and Autranella congolensis (De Wild.) A. Chev. However, the logged forest was represented by 74 records and discriminates species such as Keayodendron bridelioides (Mildbr. Ex Hutch. & Dalziel) Leandri and Guibourtia demeusei (Harms) J. Léonard;
• −In group 2, the logged forest consisted of 19 records. The characteristic species of this group were Blighia welwitschii (Hiern) Radlk, Duboscia macrocarpa Bocq, Greenwayodendron suaveolens (Engl. & Diels) Verdc. However, in group 2, the logged forest was represented by 18 sample plots, characterized by species such as Celtis zenkeri Engl., Desbordesia glaucescens (Engl.) Tiegh., Hylodendron gabunense Taub., Ongokea gore (Hua) Pierre and Uapaca guineensis Müll.Arg.;
• −Group 3 was composed of 39 records from the unlogged forest and marked by species such as Amphimas ferrugineus Pierre ex Pellegr, Klainedoxa gabonensis Pierre ex Engl, Nauclea diderrichii (De Wild. & T. Durand) Merr. and Trichilia dregeana Sond. Nevertheless, group 3 was composed of 6 records from the logged forest. The characteristic species were Alstonia boonei De Wild, Ceiba pentandra (L.) Gaertn, Funtumia elastica (P. Preuss) Stapf, Mansonia altissima (A. Chev.) A. Chev. and Triplochiton scleroxylon K. Schum;
• −In group 4, the logged forest had 15 records. It was marked by species such as Anonidium mannii (Oliv.) Engl. & Diels, Coelocaryon preussii Warb, Annickia chlorantha (Oliver) Setten & Maas, Petersianthus macrocarpus (P. Beauv.) Liben and Pycnanthus angolensis (Welw.) Warb. In contrast, the logged forest consisted of 51 concentric patches, it formed the most diverse group and was characterised by species such as Celtis adolfi-friderici Engl, Corynanthe pachyceras K. Schum. Schum, Drypetes preussii (Pax) Hutch and Pterygota macrocarpa K. Schum;
• −Group 5 was represented by 6 records from the unlogged forest. It was characterised by species such as Albizia glaberrima (Schumach. & Thonn.) Benth, Musanga cecropioides R. Br., Sterculia rhinopetala K. Schum. and Tabernaemontana crassa Benth. In contrast to the unlogged plot, the logged forest consisted of 4 sample plots. The most common species were Irvingia wombolu Vermoesen and Terminalia superba Engl. & Diels.
• −The fragmentation of the unlogged forest into more plant groups (3 plant groups) compared to the logged forest (2 plant groups) could be accounted for by the disturbance from logging activities.

Fig. 2.

Ascending classification of floristic surveys in the logged and unlogged forest.

The axes (1 and 2) of the discriminant design explain 74.79% of total inertia (Fig. 3). Axis 1 contrasts the forests according to Shannon's index, Pielou's index, Simpson's index, Fisher’s alpha index and the number of species (dbh < 10 cm and dbh >10 cm). Axis 2 shows the distribution of groups according to the average diameter (dbh < 10 cm and dbh >10 cm).

Fig. 3.

Principal Component Analysis of Shannon’s, Pielou’s, Simpson's, Alpha-fisher's indices and the number of species for the logged and unlogged forests.

3.2. Alpha diversity

3.2.1. Taxonomic and species diversity

The flora of the unlogged forest was constituted of 171 woody species, distributed in 120 genera and 56 families for individuals of dbh≥10. The understorey (individuals of dbh<10 cm) was constituted of 48 species, 45 genera and 23 families. In the logged forest, the species richness was 126 woody species, 102 genera and 37 families for individuals of dbh≥10 while the understorey was composed of 44 species, 41 genera and 25 families. Comparing the two forest types inventoried revealed higher diversity of species, genera and families for the unlogged forest (Table 1). Similar trends were observed for the diversity indices, except for Pielou's equitability index, which showed higher representativeness of species in the logged forest compared to the unlogged forest.

Table 1.

Floristic richness and diversity in the logged and unlogged forests. *p < 0.05.

						Confidence interval at 95%
Diameter of species		Unlogged	Logged	Mean	p-value	Superior limit	Inferior limit
dbh≥10 cm	Taxonomic richness
	Species	171	126	148 ± 31.82	0.65	193.5	103.5
	Genera	120	103	111 ± 12.02	0.38	128.5	94.5
	Families	48	37	42 ± 7.77	0.31	59.5	37.5
	Species diversity
	Shannon	4	3.86	3.92 ± 0.09	0.58	4.05	3.79
	Simpson	0.96	0.94	0.95 ± 0.01	0.5	0.97	0.93
	Alpha-fisher	43.18	27.8	35.49 ± 10.87	0.02*	50.87	20.11
	Pielou	0.77	0.83	0.8 ± 0.04	0.35	0.86	0.74
dbh<10 cm	Taxonomic richness
	Species	48	44	46	0.17	71.41	20.59
	Genera	45	41	43	0.31	68.41	17.58
	Families	23	25	24	0.25	36.71	11.29
	Species diversity
	Shannon	2.67	2.8	2.73	0.54	3.56	1.91
	Simpson	0.86	0.88	0.87	0.36	0.99	0.74
	Alpha-fisher	13.33	18.42	15.87	0.84	48.21	16.47
	Pielou	0.68	0.72	0.7	0.17	0.44	0.95

3.3. Rarefaction curve

The accumulation curve (Fig. 4) shows a variation in the number of species between 1 and 171 species. However, it is important to note that the sampling effort in the logged and unlogged forests was the same. Nevertheless, there was a considerable sampling effort overall. The increasing trend of the curve in the logged forest indicates the appearance of new species as the number of individual increase. On the other hand, in the logged forest, the evolution of the curve indicates the peak of sampling of species when the number of individuals increases.

Fig. 4.

Rarefaction curve.

3.4. Characteristic species

For the two forest types studied, a list of species with an importance value index greater than 10 was drawn up. There were 11 species in the unlogged forest and 8 in the logged forest (Fig. 5). In ascending order, the most ecologically important species in the unlogged forest were Uapaca guineensis Müll.Arg. (89.6%) while in the logged forest, the most important species was Triplochiton scleroxylon K. Schum. (47.6%). The overall IVI was 18.32 ± 0.03 in the unlogged forest and 5.49 ± 0.02 in the logged forest (p-value = 0.0001). This disparity in value reveals that logging influences the distribution, representativeness and dominance of individuals of a species. The statistical test between the IVIs of the important species in the different plots shows that there is no significant difference (p-value = 0.386 in the unlogged forest and p-value = 0.054 in the logged forest).

Fig. 5.

Importance Value Index of the most dominant species. IVI: Importance Value Index Uap_gui: Uapaca guinensis, Tri_scl: Triplochyton scleroxylon, Tri_dre: Trichilia dregeana, Tab_mon: Tabernaemontana crassa, Ste_rhi: Sterculia rhinopetala, Pyc_ang: Pycnanthus angolensis, Pol_sua: Polyalthia suaveolens, Mus_cec: Musanga cecropioides, Man_alt: Mansonia altissima, Ena_chl: Enantia chlorantha, Dub_mac: Duboscia macrocarpa, Des_gla: Desbordesia glaucescens, Cei_pen: Ceiba pentandra, Bli_wel: Blighia welwitschii, An_man: Anonidium mannii.

3.5. Correlation between basal area and stem density

Fig. 6 shows the relationship between basal area and stem density in the logged and unlogged forests. The basal area and stem density showed a negative non-linear relationship in both forests. The correlation was weak in both forests (R2 = 0.386 in the unlogged forest and R2 = 0.355 in the logged forest). The two variables were significantly different (p-value <0.001).

Fig. 6.

Correlation between basal area and stem density in the logged and unlogged forests. ***p < 0.001.

3.6. Characteristic genera

The generic coefficients (0.51 ≤ GC ≤ 6.84) foro the logged and unlogged forest were relatively low (Table 2). In the unlogged forest, the most diverse genera were Celtis (4 species: C. adolfi-friderici Engl., C. africana Burm. f., C. tessmannii Rendle, C. zenkeri Engl.), Irvingia (4 species: I. excelsa Mildbr., I. wombolu Vermoesen, I. grandifolia (Engl.) Engl., I. robur Mildbr) and Xylopia (4 species: X. hypolampra Mildbr. & Diels, X. phloiodora Mildbr, X. quintasii Pierre ex Engl. & Diels, X. staudtii Engl. & Diels). The number of genera was higher in the unlogged forest and the number of species per genus decreased in the logged forest. The most diversified genera in the logged forest had at most three species. These were the genera Celtis (3 species: C. adolfi-friderici Engl., C. africana Burm. f., C. zenkeri Engl.), Irvingia (3 species: I. wombolu Vermoesen, I. grandifolia (Engl.) Engl., I. robur Mildbr), Cola (3 species: C. altissima Engl., C. argentea Mast., C. ballayi Cornu ex Heckel)). The high number of genera in the unlogged forest compared to the number of genera in the logged forest indicates the loss in plant diversity due to logging activities. However, statistical tests show that there is no significant difference between the different genera in the two forest types (p-value >0.05).

Table 2.

Generic coefficient as a function of the number of species in the forest type. NG: Number of genera; GC: Generic coefficients.

	Forest type
	Logged		Unlogged		Statistical test
Number of species	NG	GC (%)	NG	GC (%)
1	91	6.84	95	4.08	X2 = 2.26; df = 2; p-value = 0.32
2	10	1.50	14	1.20
3	3	0.68	8	1.03
4	–	–	3	0.51
Total	103	9.02	120	6.85

3.7. Abundance of genera per family and Family Importance Value

Regarding the number of genera per family in the whole massif (Fig. 7), the values varied from 1 to 18 and the most diversified in decreasing order were: Fabaceae (17 genera), Malvaceae (13 genera), Annonaceae (8 genera), Euphorbiaceae (8 genera), Sapotaceae (7 genera), Cannabaceae, Meliaceae, Rubiaceae and Strombosiaceae, with 5 genera each in the unlogged forest. In the logged forest, there was similarity between Fabaceae (17 genera), Annonaceae (8 genera), Meliaceae (5 genera) and Rubiaceae (5 genera) with the unlogged forest. In addition, the result revealed a lower number of genera in certain families for the logged forest compare to those of the unlogged forest. These include Malvaceae (8 genera) and Euphorbiaceae (6 genera).

Fig. 7.

Characteristic families. FIV: Family Importance Value.

The Family Importance Value (FIV) analysis showed that 11 families were more ecologically important (FIV10) in the Dimako Communal Forest. The Malvaceae (34.16) family was more important in the unlogged forest, while in the logged forest, the Fabaceae (30.8) family was more important. In relation to the important families in the two sites, we noted the individualisation of Moraceae and Euphorbiaceae in the logged forest. However, in the unlogged forest, only Meliaceae were the main characteristic families of undisturbed forest environments. In relation to the important families in both sites, we noted the individualisation of Moraceae, Euphorbiaceae and Malvaceae in the logged forest. However, in the unlogged forest only Myristicaceae was the main characteristic family. The overall family importance was higher in the logged forest (36.12 ± 0.11) than in the unlogged forest (35.07 ± 0.05). FIV was high in the logged forest (36.12 ± 0.11) and lower in the unlogged forest (35.07 ± 0.05). The Kruskal-Wallis statistical test showed a significant difference (p-value<0.001) between the logged and the unlogged forests.

3.8. Beta diversity

From the point of view of diversity, Sørensen's similarity index showed a high floristic similarity (Sørensen = 71.49%) between the unlogged and the logged forests.

The floristic background and cover were dominated by Guinean-Congolese endemic species for both forests. The proportions were 67.57% and 63.07%, respectively in the unlogged and the logged forests. Binding species were poorly represented with 2.64% and 2.87%, respectively in the unlogged and logged forests.

3.9. Life spectra of woody species in the plots

The analysis of the phytogeographical distribution showed ten phytogeographical domains in the two forest types. The unlogged forest differed from the logged forest in that the Guineo-Congolese-Zambesian and the Lower Guineo-Coastal domains were present with 100% representation. The indicator species of these domains were Entandrophragma angolense C. DC, Isolona hexaloba (Pierre) Engl. & Diels, Tetrapleura tetraptera (Schumach. & Thonn.) Taub. and Markhamia lutea (Benth.) K. Schum. On the other hand, the logged forest was characterized by the presence of the Guinean-Congolese-Sudanese domain. The indicator species was Balanites wilsoniana Dawe & Sprague. These results showed that in the unlogged forest, the edaphic stationary conditions were favourable to the emergence of the Guineo-Congolese-Sudanese flora.

The analysis of the species contingent showed the existence of six biological types with a preponderance of Phanerophytes. The unlogged forest was distinguished from the logged forest by the predominance of Macrophanerophytes (100%), with characteristic species such as Anthonotha lamprophylla (Harms) J. Léonard, Baillonella toxisperma Pierre, Barteria fistulosa Mast, Ceiba pentandra (L.) Gaertn, Celtis tessmannii Rendle, Detarium macrocarpum Harms, Diospyros crassiflora Hiern, Funtumia elastica (P. Preuss) Stapf, Garcinia mannii Oliv, Keayodendron bridelioides (Mildbr. Ex Hutch. & Dalziel) Leandri and Macaranga hurifolia Beille (Fig. 8). The abundance of the latter proves the dominance of the middle stratum in the studied forest types.

Fig. 8.

Functional spectra of species. Legend. Pal: Paleotropical; At: Afro-tropical; BG: Lower-Guinean; BGC: Lower-Guineo-Congolese; BG-Cot: Lower-Guineo-Coastal; G: Guinean; GC: Guineo-Congolese; C–Z: Congolo-Zambian; GC-Z: Guineo-Congolese-Zambian; GC-S: Guinea-Congo-Sudanese; Mgph: Megaphanerophytes; Mph: Macrophanerophytes; Msph: Mesophanerophytes; Mcph: Microphanerophytes; Nph: Nanophanerophytes; Bar: barochores; Scl: sclerochores; Bal: ballochores, Pog: pogonochores; Pte: pterochores; Sar: saccochores; Se: sedentary species; Ci: species with ephemeral scarring temperament; Pi: pioneer species.

In both forests surveyed, eight types of diaspores were identified. Heterochores constituted the bulk of the diaspores in the Dimako Communal Forest with their proportions ranging from 86.99% in the logged plot to 90.48% in the unlogged forest (p-value = 0.04). This large group was dominated by sarcochorous species which were strongly represented in the unlogged forest (75.5%) compared to the logged forest (24.5%). The group of autochores was characterised by the abundance of sclerochores and ballochores in the unlogged forest. The low representativeness of sclerochores and ballochores in the logged forest is an indication of the presence of an area disturbed by industrial logging activities.

The flora of the three temperament types was recorded in both forests. The analysis reveals that sedentary species (73.5%) were the most dominant in the unlogged forest. The characteristic species for this biological type were Allanblackia floribunda Oliv, Antiaris toxicaria var. africana Scott-Elliot ex A. Chev, Blighia welwitschii (Hiern) Radlk, Diospyros bipindensis Gürke, Leplaea thompsonii (Sprague & Hutch.) E. J. M. Koenen & J. J. de Wilde, Santiria trimera (Oliv.) Aubrév. and Trichoscypha acuminata Engl. These are the species that contribute to the construction of the climatic stage. However, in the logged forest, pioneer species were predominant (34.6%). These species represent the group that first colonises disturbed environments. This group was composed of species such as Cleistopholis patens Engl. & Diels, Erythroxylum mannii Oliv, Lannea welwitschii (Hiern) Engl, Myrianthus arboreus P. Beauv, Pteleopsis hylodendron Mildbr, Ricinodendron heudelotii (Baill.) Pierre ex Heckel, Trema orientalis (L.) Blume, and Zanthoxylum heitzii (Aubrév. & Pellegr.) P. G. Waterman. Statistical tests revealed that there was no significant difference (F = 0.02, df = 5.99, p-value = 0.088) between the temperament spectra in the two forest types.

4. Discussion

The factors determining the spatial distribution of plant groups are environmental gradients such as topography on the one hand, and intensity of land use and soil cover on the other. These findings are similar to those of [47] who showed that the grouping of species seems to be defined by edaphic conditions and local geomorphology. Each plant grouping is characterised by its specific edaphic conditions and indicator species that strongly contribute to the maintenance of the grouping despite disturbances. According to Ref. [48], the extinction or significant fluctuation in the abundance of indicator species can seriously affect other species and lead to or accelerate the extinction of the entire plant community. Furthermore, indicator species can be used to analyse vegetation trends and their underlying environmental variables [49]. Also, the significant influence of edaphic factors and relief in the spatial distribution of plant groups within the natural landscape had also been highlighted by Ref. [47].

In ecology, diversity indices are the best parameters for characterising a stand [50]. We note a great diversity of species and genus in the unlogged forest compared to the logged forest. This is confirmed by the variation in the value of the Shannon index (4 bits for the unlogged forest compared to 3.86 bits for the logged forest for stems of dbh 10 cm). The low species richness in the logged forest could be due to the removal of valuable species for commercial purpose or by their accidental destruction through logging activities and the early disappearance of pioneer species. This observation was made by Ref. [51] in the forest plantations of Mangombe. According to this author, in disturbed environments, pioneer species were not firmly and durably established in the communities where they appeared. The logged forest was characterised by a regular distribution of individuals within the species, as shown by the high value of Pielou's equitability (0.83) in the logged forest compared to the unlogged forest (0.77). These findings are contrary to those of [25,48,52] who showed that in less disturbed sites, species were more evenly distributed. These results are similar to those of [4,53] who stated that logging leads to successional changes in floristic diversity. Furthermore [54], showed that forest maturity can lead to a decrease or increase in biodiversity. Despite the fact that these two forest types are rich, the unlogged forest stands out for its high heterogeneity, which is explained by the high value of the Fisher-alpha index (43.18). This could be explained by its status (unlogged), the abundance of sedentary species, and Mesophanerophytes that contribute to the closure of the canopy. These observations are similar to those of [55], who stated that a moderate disturbance favours an increase in floristic diversity. This explains the high heterogeneity observed in the unlogged forest.

Owing to the forest management applied in the Dimako Communal Forest, the results of this study suggest that low intensity industrial logging has little negative impact on the floristic diversity and species richness for species of dbh <10 cm individuals. These results also suggest that secondary succession is more favourable for the reconstitution of forests that have undergone low intensity logging. This is justified by the number of species (48 species in the unlogged forest compared to 44 species in the logged forest) and the value of the Shannon diversity index (2.67 bits in the unlogged forest and 2.8 bits in the logged forest). The high similarity between these two forest types is observed in the Sørensen similarity index (Ks 50%), which confers on them membership to the same plant community. However, the differences between the logged and unlogged forests are only apparent, as the two forest types are not significantly different in taxon richness and species diversity (p-value <0.5). Although industrial logging has caused the loss of some taxa in the logged forest, the impact remains small. This is because the intensity of tree removal was low, about 0.78 trees per hectare [29] and also the time after the industrial logging (15 years) allowed the logged forest to reconstitute the lost vegetation.

Regarding families, Malvaceae (former Sterculiaceae, Bombacaceae and Tiliaceae according to the APG IV classification) were predominant in the unlogged forest compared to the Fabaceae family (former Caesalpinianceae, Papilionaceae and Mimosaceae), predominant in the logged forest. These results are similar to those obtained by some authors [31,42,56,57] who stated that the semi-deciduous dense humid forest is characterised by the abundance of Sterculiaceae (included in Malvaceae today). The abundance of Fabaceae in African dense forests is one of the fundamental characteristics that differentiates them from those of Asia and brings them closer to those of the Americas [45].

Species with a large distribution were more represented in the unlogged forest (4.16%%) compared to the logged forest (3.91%). According to Ref. [58], this difference illustrates the secondary character of the logged forest and therefore the absence or low anthropisation of the unlogged forest. However, regardless of the high anthropogenic pressure (logging) in the logged forest, the flora has not completely lost its specificity. This conservatory state observed in the high proportion of Guinean-Congolese species (67.57%) in the unlogged forest compared to the logged forest (63.07%) confirms the belonging of the study area to the dense humid forest zone as defined by White (1986). The values obtained were close to those mentioned by some authors who worked in the dense rainforest zone [42,57,59,60]. However, the percentages found are in line with the predictions made by Ref. [45]. Indeed, the latter considers the Guinean-Congolese flora to be remarkably pure with more than 80–90% endemism and only 10% of linkage species.

Biological type is far from being one of the important indices that illustrate the vertical structure, the physiognomy of plant formations and the degree of disturbance of an environment. As reported by Ref. [61], the vertical profile provides information on stand height, degree of closure, density, crown spread, vegetation tiering from the ground to the canopy. However, the biological spectrum of all the forests sampled was marked by the dominance of the Phanerophyte group, particularly the Mesophanerophytes. The latter represent the dominant subset of this group. The high frequency of Mesophanerophytes in the logged forest could be due to the fact that, in previously disturbed forests, the tree crowns are not joined and the opening of the canopy caused by human activities favours their extension. These results corroborate those of [51,62] who reported that Mesophanerophytes are the main characteristic species of disturbed environments, due to the absence of large trees. The importance of Mesophanerophytes in plant succession has also been highlighted by Ref. [59] in the Bamo classified forest in Côte d'Ivoire [63], in the Gracinia spp. transitional forests of the Nkilobot hills in Cameroon, and [64] in the Mount Cameroon line massif (Mount Kupe).

In both forest types studied, the preponderance of heterochorous species was noted. This predominance can be explained by the constant flow of species regardless of age and type of land use. These observations mean that some types of diaspora are more subject to species flux than others through colonisation, extinction and dispersal mechanisms [65]. mentioned that the presence of diaspores determines the composition of the first biocenoses of the succession.

The predominance of heterochores observed in this study corroborates with data obtained in the forests of the Congo Basin and confirms zoochory, particularly endozoochory as the main mode of dissemination in the Dimako Communal Forest. This dominance of endozoochory reflects the intense activity of frugivorous/disseminating animals (rodents and birds) and is a positive impact of the dynamics, although logging leads to the displacement of some animals, hence the low spread in disturbed areas. These results are similar to those of [57,66,67] who worked in dense rainforests. Thus, it can be emphasised that the existence of a good balance between vegetation cover and wildlife is a guarantee for the sustainability of forests. However, the spectrum of diaspore types in this forest reveals the predominance of plants with drupaceous fruits or arillate seeds. These results are similar to those of some predecessors [57,64,68] who have shown that a significant number of animals depend on fleshy-fruited species for their food. As a result of the disturbances noted, a fairly high proportion of pterochorous taxa (on average 0.97%) was noted in the logged forest.

The temperament spectrum in the both the logged and unlogged forests studied showed significant proportions of sedentary species. This reveals that low logging intensity allows the forest to maintain its characteristics of dense humid forest. This result also suggest that low logging intensity logging allows for reconstitution of the forest under controlled conditions after logging, that is, when human disturbance is limited. The preponderance of pioneer species was noted in the logged forest. The increase in sedentary species densifies the vegetation cover by reducing the intensity of light in the soil and promoting the decomposition of the litter. This decomposition stimulates the germination of pioneer species in the environment. These findings are similar to those of [69] who showed that increasing soil temperature causes the rapid decomposition of soil humus. The identification of indicator species in each stand shows that the two forest types are indeed different in terms of species temperament. Species with an ephemeral scarring temperament were less represented in the unlogged forest and more abundant in the logged forest. The relatively low proportion of these species indicates that the unlogged forest is the result of an old disturbance (15 years for the logged forest surveyed in this study).

5. Conclusion

The present study was carried out in the dense semi-deciduous forest of Cameroon, particularly in the Dimako Communal Forest. The objective of this study was to provide comparative floristic knowledge that can serve as a basis for the planning and sustainable management of ligneous plant resources in Communal Forests before and after logging. Biocenotic indices showed that the unlogged forest was more diverse than the logged forest. This representative diversity is related to the status of this forest type (unlogged). Sorensen's similarity index showed that the two forest types form the same ecological grouping, despite anthropogenic disturbances and particularly logging activities. In both forest types studied, the results showed that despite anthropogenic action, the majority of the flora is made up of Phanerophytes, particularly Mesophanerophytes. The phytogeographical distribution of the flora of the two forest types is related to the Guinean-Congolese centre of endemism. In order to promote the sustainable management of the Communal Forest in Cameroon, it is essential to carry out such analyses over time and space. This findings of this study reveal that the synchronous study of woody diversity is an important step in establishing a basis for sustainable management of production forests as it provides evidence of the past flora, reflects current plant diversity and enables management planning that takes into consideration future generations. Notwithstanding the changes in the DCF between the two periods, it remains largely covered by relatively well-preserved vegetation and has good potential for biodiversity conservation. The implementation of forest management practices in this massif has contributed to the reconstitution of the vegetation. This leads us to conclude that low intensity industrial logging and secondary succession favour the reconstitution of logged forest under controlled anthropogenic action after logging meanwhile assisted natural regeneration is needed to facilitate the reconstitution of the taxa lost due to logging activities.

Author contribution statement

Hubert Mounmemi Kpoumie: Conceived and designed the experiments; Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Ekué Marius Rodrigue Mensah: Performed the experiments; Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.

Forbi Preasious Funwi: Analyzed and interpreted the data; Contributed reagents, materials, analysis tools or data.

Louis Paul Roger Kabelong Banoho: Analyzed and interpreted the data; Wrote the paper.

Bertine Tiokeng: Maffo Maffo Nicole Liliane: Ntonmen Yonkeu Amandine Flore: Evariste Hermann Taedoumg: Wrote the paper.

Betti Jean Lagarde: Tchoupou Votio Carole Mireil: Contributed reagents, materials, analysis tools or data; Wrote the paper.

Zapfack Louis: Performed the experiments; Contributed reagents, materials, analysis tools or data; Wrote the paper.

Data availability statement

No data was used for the research described in the article.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.