This study was conducted in the community forests of Nkala and Embirrima, located in the Guineo-Congolian zone in the southwest of the Mai Ndombe province ([5]), Bolobo territory, and which are part of the six community forests grouped under the label of the “Community Forest Concessions of the Mbali River” (CFCL-RM) ([33]). The choice of the Nkala and Embirrima forests is explained in that they are the oldest to have been ceded for the conservation of the Bonobo habitat. These two forests have substrates on hydromorphic soils, periodically flooded, and firm lands ([36]; [37]).

The fauna is diverse with a dominance of bonobos (Pan paniscus) as the main disperser and a strong presence of birds and other species ([5]; [38]). The local Téké population living there are practicing slash-and-burn agriculture and are highly dependent on the forest for subsistence farming. Households in our area clear only half a hectare per year on average and rarely more than one hectare ([9]).

2.2.1. Field Data Collection

“Pre-forest fallow” is defined in this study as a postagricultural site that has been abandoned after slashandburn cultivation and is undergoing natural secondary succession toward a forest state. Only fallows ≥ 10 years since abandonment were included, ensuring that the initial herbaceous and early shrub phases had been surpassed. Fallow age for each plot was determined through semi-structured interviews with local land chiefs and the farmers who last cultivated the plot, who provided the year of final clearing and cropping. The number of years since abandonment was calculated from that date, and plots were subsequently assigned to six age classes (10 - 19, 20 - 29, 30 - 39, 40 - 49, 50 - 59, and ≥ 60 years). Within each fallow land, we established a 0.25 ha plot, following established clearing practices ([17]; [22]; [29]). Each plot was established from the center outwards to minimize the risk of introducing bias as shown in Figure 1. A total of 20 square plots in fallows ≥ 10 years since abandonment were established across two sites: 10 plots in Nkala and 10 in Embirrima.

Figure 1.Data collection plot.

Once the plot was established, the next step was to take the geographic coordinates from the center. We identified all species of trees, shrubs, seedlings, herbs, and lianas, and measured the heights of the trees and shrubs, as well as the DBH (Depth of Water) of species ≥ 10 cm. The plants were classified into strata according to [4].

2.2.2. Species Identification

Species identification was made possible using several identification keys and botanical indicators. The samples were compared at the botany laboratory of the Department of Biology and at the Laboratory of Systemic Biodiversity, Conservation, and Endogenous Knowledge of the Department of Environmental Sciences and Management at the University of Kinshasa. We also used several illustrative and descriptive documents on tropical species worldwide, primarily in Central Africa, available both online and in print.

Canopy opening was estimated visually as the percentage of canopy cover per 50 × 50 m plot. Understory density was quantified as the combined cover percentage or stem count of all vegetation in the shrub (1 - 5 m) and herbaceous (≤1 m) strata, also recorded per plot. Pearson correlation coefficient (r) was calculated to assess the relationship between canopy opening (%) and understory density (%). The test was performed using R software (version 3.3.1).

2.3.1. Autoecological Characteristics

The various intrinsic characteristics of the species were determined by observation according to the approach proposed by [8], [40], and [43]; and subsequently adopted by [21], [26], [4], [7], and [19]. The elements considered were diaspore type (TD), biological type (TB), leaf type (TF), phytogeographic distribution (DP), and phytosociological status (TP). The species form was also taken into account.

2.3.2. Degree of Fallow Disturbance

We determined the degree of fallow disturbance using the [45] formula, as employed by [7]. By assessing the relative importance of pioneer species compared to all species inventoried at this site, the “Pioneer” index was calculated using the following formula:

PI = [(pi + np)/N] × 100

where pi is the number of individuals of pioneer species, np the number of individuals of non-pioneer but sun-loving species, and N the total number of individuals of all inventoried species. PI ranges from 0 (absence of pioneer species and therefore of disturbance) to 100 (all species are pioneers; the forest is completely disturbed and degraded). This Pioneer Index is a good indicator of the degree of disturbance or degradation of the forest.

Diversity and similarity indices such as the Shannon, Pielou, Simpson, Jaccard, and Sorensen indices were used to analyze the biodiversity between the two vegetations of the CFCL-RM. The Shannon-Weaver index (H') allows for measuring diversity by taking into account both species richness (number of species) and evenness (distribution of abundances among species).

A high value indicates a diverse and balanced community (many species, none very dominant). A decrease may indicate environmental stress. Much more sensitive to the presence of rare species than the Simpson index, it is the most commonly used in community ecology. It allows for solid comparisons between different sites or periods.

Interpretation: 0 - 1 = low, 1 - 3 = medium, >3 = high.

Pielou’s evenness (J') measures the uniformity of abundance distribution among the species present. It is the ratio between observed diversity (H') and the maximum possible diversity (H'max = ln(S)). It complements the Shannon index by addressing whether diversity is low due to a lack of species or very strong dominance. A value close to 1 indicates that all species have the same abundance. Close to 0, one or a few species dominate. It is key to interpreting changes: A decrease in H' with a stable J' indicates a loss of species. A decrease in H' with a drop in J' indicates an increase in dominance.

Simpson (D): measures the probability that two individuals randomly drawn from the community belong to the same species. The original index (D) is a measure of dominance. 1 − D or 1/D is often used as a measure of diversity. It emphasizes dominant species: It is more sensitive to changes in the abundance of common species. It is useful for assessing resistance to invasions. A high dominance (high D) can sometimes indicate a less resilient community.

Sørensen (Sø) and Jaccard (J) measure the compositional similarity between two communities (beta diversity). They are based on the presence/absence of species. These two indices prioritize conservation by identifying sites that harbor unique communities (low similarity) and thus complementary for a network of reserves. However, while Sorensen gives more weight to common species than to unique ones, Jaccard treats them equally. Sø is generally preferred in ecology. In our work, we used both formulas.

Sorensen C S = 2 a 2 a + b + c

Jaccard C J = a a + b + c

The inventories conducted in the forest fallows of Nkala and Embirrima on dry land in the CFCL-RM reported 2,578 individuals belonging to 170 species, 126 genera, and 59 families (APGIII). The most representative family’s number 5, including the Fabaceae group with 12% (Mimosoideae 3%; Faboideae 4% and Caesalpinioideae 5%), followed by Annonaceae and Apocynaceae with 8% each, then Euphorbiaceae with 7%, and finally Rubiaceae with 5%.

The other families are poorly represented. The vegetation of Nkala (135) presents a higher number of taxa than that of Embirrima (114) as shown in Table 1. Thus, the two sites are not significantly different with p-Value (0.227) > 0.005.

Table 1. Spatial variation of different biodiversity indices of pre-forest fallows.

The Simpson_1-D index indicates that both sites have high species diversity (values close to 1, meaning low dominance). Contrary to the original statement, the Shannon index is slightly higher in Nkala (H' = 4.03) than in Embirrima (H' = 3.88), suggesting marginally greater diversity in Nkala, while Simpson values are nearly identical (0.9593 and 0.9599), confirming that overall diversity is comparable between the two sites. The evenness index J ′ (corrected) is approximately 0.82 for both sites, indicating a balanced distribution of species abundances. The Jaccard and Sørensen similarity indices (0.33 and 0.49, respectively) show moderate floristic similarity between Nkala and Embirrima.

The biological traits considered in this study concern the geographical distribution of species, the types of diaspores, the biological types, and the leaf types.

3.2.1. Phytogeographic Spectrum

The proportions of the geographical distribution of the collected species (Figure 2) highlight the predominance of Guineo-Congolian species (GC) (47.0%). They are followed by the species of the Lower Guineo-Congolian (20%), and finally the Afro-Tropical (14%). Practically no endemic species have been identified in the fallow land flora.

Figure 2. Phytogeographic spectrum of fallow land.

3.2.2. Diaspore Types

A detailed examination of diaspore types (Figure 3) shows a large predominance of species with fleshy diaspores (Sarcochores: 68%) in our study. These are diaspores from shrubs, trees, and overstory vegetation. Ballochorous species (19%) are the second most common, followed by Pterochorous species (6.4%).

Figure 3.Spectrum of diaspore types in the fallow flora.

3.2.3. Biological Types

The biological types (Figure 4) most represented in the flora of the studied fallows are the Mesophanerophytes (MsPh) with 74%, followed by the Lianous Phanerophytes (Lph) with 23% and the Megaphanerophytes (MgPh) with 23%. Phanerophytes are dominant in the pre-forest fallows of the CFCL-RM, however, a chamaephyte tendency is very weakly observed.

Figure 4. Biological spectrum of fallow lands in CFCL-RM.

3.2.4. Leaf Types

Figure 5. Analysis of the leaf spectrum of the CFCL-RM fallows.

Analysis of leaf size types (Figure 5) indicates a dominance of mesophyll species (75%). Microphyll species (10%) are the second most common. The other categories are poorly represented. Mesophyll species are largely dominant in the different fallow areas.

The Pioneer Index (PI), which gives us an image of the degradation state of the two sites (Figure 6), is generally low overall with an average PI of ± 45%. This value indicates that, overall, the forest stands are slightly disturbed. Individually considered, Nkala is the most disturbed site with a PI of 55% and 45% for Embirrima.

NP = non-pioneer species; Pi = pioneer species; SB = shade-tolerant species.

Figure 6. Raw spectra of temperament and pioneer index by site.

AcD: Co-Dominant Trees; Ar: Shrubs; AD: Dominant Trees; AC: Canopy Trees, L: Vines; Pl: Seedlings.

Figure 7.Variation in the number of individuals based on morphologies by age class in the pre-forest fallows of the CFCL-RM.

Plant morphology variation by age showed a decreasing curve in the number of trees between 30 years and 60 years as shown in Figure 7. Dominant trees (DT) are much more represented in increasingly older fallows with more than 50 trees between 10 years and 60 years. We observe a decline in canopy trees (CT) and seedlings (S) with age.

The correlation between canopy opening and understory density was strong and statistically significant (r = 0.82, p < 0.001, n = 20), confirming that understory density increases with greater canopy openness as shown in Figure 8. The highest understory density occurred in the 30 - 39 year fallow class, where canopy opening peaked.

Figure 8.Variation of the undergrowth depending on the canopy opening in the fallows of the CFCL-RM.

The floristic inventory revealed a notable diversity in the pre-forest fallows of the CFCL-RM, with 170 species distributed across 59 families. This richness, although lesser than that of the neighboring primary forests ([5]), is significant for post-agricultural regeneration ecosystems. The superiority of the number of taxa at Nkala (135) compared to Embirrima (114) could be explained by the nature of the landscape: Nkala constitutes a continuous forest block, potentially more favorable for the accumulation of species than the forest-savanna mosaic of Embirrima, where ecotones and different historical disturbances can filter certain species ([34]). The Shannon diversity indices (H' ≈ 4.0 and 3.9) and Simpson (1 − D > 0.95) indicate structured and diverse communities, typical of African secondary forests in the process of reconstitution ([27]; [14]). The evenness values (J' ≈ 0.82 for both sites) suggest a relatively balanced distribution of abundances among species, indicating intra- and interspecific competition not yet dominated by a few taxa, characteristic of an intermediate stage of succession.

4.2.1. Phytogeographic Spectrum

The strong predominance of species from the Guineo-Congolian domain (47%) and the Lower Guineo-Congolian sector (20%) confirms the biogeographical affiliation of the flora studied in the heart of the Congo basin. The almost total absence of strict endemics reflects the secondary and still young nature of these fallows, where the flora is mainly composed of species with a wide distribution and good dispersal capacity, typical of post-disturbance stages ([13]). The presence of Afro-tropical elements (14%) indicates the influence of savanna corridors or more open environments, particularly in Embirrima.

4.2.2. Types of Diaspores

The overwhelming dominance of sarcochores (fleshy diaspores, 68%) is a major result. It highlights the crucial role of zoochory, and particularly frugivory, in the recruitment and regeneration of these fallows. This predominance is consistent with the presence of a dispersal fauna, particularly bonobos (Pan paniscus) identified as key players in the landscape ([5]; [38]). In fact, Based on the combination of diaspore syndromes and prior studies on bonobo ecology in the same landscape, a plausible inference can be made that bonobos actively shape forest regeneration, even though direct dispersal observations were not conducted in this study.

The inventoried flora shows a clear predominance of sarcochores (fleshy diaspores), a dispersal syndrome directly associated with frugivorous animals. This structural trait alone confirms that zoochory is the dominant regeneration mechanism. Prior research in the Mbali landscape ([38]; [34]; [5]) has documented bonobos (Pan paniscus) as the largest and most abundant frugivores, with diets composed largely of the same tree species that dominate these fallows. Moreover, the spatial continuity of the forest block and the cultural prohibitions against bonobo hunting in the study area further support their role as effective, unhindered dispersers.

Therefore, while this study did not directly measure seed ingestion, gut passage, or dispersal distances, the convergence of functional trait analysis (sarcochore dominance) and existing ecological knowledge (bonobo frugivory and habitat use) provides strong confirmatory evidence that bonobos are key agents in the successional dynamics and forest evolution of the CFCLRM. Their frugivory likely drives the recruitment of many late-successional species, linking faunal conservation directly to forest recovery and carbon sequestration.

The low representation of pterochorous plants (6%) suggests that anemochory plays a secondary role in the colonization of these environments, perhaps due to the closed structure that young fallows quickly develop.

4.2.3. Biological Types

The community is largely dominated by phanerophytes, a category that includes nearly 97% of the species, with a clear predominance of Mesophanerophytes (74%). This spectrum is characteristic of a forest formation undergoing vertical structuring. The high proportion of liana phanerophytes (23%) is a classic indicator of recent disturbance and the exploitation of available light in the early stages of succession ([22]). The scarcity of chamaephytes and therophytes confirms that the environment has surpassed the initial stage of pioneer herbaceous plants to evolve into a structured woody population.

4.2.4. Leaf Types

The clear dominance of mesophylls (75%) is a functional trait adapted to moderate light conditions and high atmospheric humidity under a closed forest canopy. This result indicates that, despite their post-cultural origin, the studied fallows have recreated an internal forest microclimate that favors this type of foliage, to the detriment of microphylls (10%) which are more suited to conditions of greater water or light stress ([4]).

The average Pioneer Index (PI) of 45% confirms that the stands are in an advanced phase of succession, far from a pure pioneer stage. The difference between Nkala (PI = 55%) and Embirrima (PI = 45%) is significant. The Nkala site, although a forest block, has a slightly younger or more frequently disturbed stand (probably due to local human activities such as collection), favoring a higher proportion of heliophilous species. Embirrima, despite its mosaic landscape, harbors slightly more mature or less disturbed stands, with a higher proportion of shade-tolerant species. This nuance highlights that the history of use and the configuration of the landscape finely influence the successional trajectory.

The morphological analysis reveals an expected but instructive demographic dynamic. The decreasing curve of the total number of individuals with age, associated with the increasing proportion of dominant trees (AD), illustrates the process of self-thinning and competition for light. The scarcity of seedlings (Pl) and canopy trees (AC) in the oldest classes suggests that regeneration under closed canopy is becoming limiting. This indicates a possible bottleneck for the recruitment of shade-loving species in the absence of gap-creating disturbances, a key natural process for maintaining long-term diversity ([14]; [19]).

The strong positive correlation observed between canopy opening and understory density is a cornerstone of tropical forest ecology, clearly confirmed in our context. The peak in understory density in the 30 - 39 year age class likely corresponds to a phase where the canopy of fast-growing pioneer species begins to open (through senescence or competition), allowing for a strong surge in sun-loving regeneration and liana growth. Subsequently, the progressive closure of the canopy by longer-lived, broad-crown species leads to understory thinning, as observed in older fallows. This “wave” dynamic is characteristic of secondary succession and underscores the importance of gap cycles for forest renewal and stratification ([12]; [22]).

Overall, this study demonstrates that post-AIB fallows in CFCL-RM, beyond 10 years, are dynamic and complex ecosystems engaged in an active process of forest regeneration. The flora is already dominated by traits of mature forest (Guineo-Congolian phytogeography, dominance of mesophyll sarcochores), but the demographic structure and temperament indices reveal that they are still in an intermediate phase of succession. Comparison between sites shows that the successional trajectory is modulated by the landscape context and the intensity of residual disturbances. The strong dependence on zoochory highlights the critical interdependence between forest regeneration and the conservation of frugivorous fauna, particularly bonobos. These secondary forests therefore play an irreplaceable role as a biodiversity reservoir and ecological corridor in a human-modified landscape, fully justifying their integration into conservation and land-use planning strategies ([30]; [41]).