Woody species diversity and dispersal dynamics in Biyo Forest, Central Ethiopia: a conservation perspective

Abstract

Background

Seed dispersal shapes species composition, regeneration, and long-term resilience of dry Afromontane forests. In Ethiopia, there is limited information on how woody species disperse and recruit. Biyo Forest, a protected exclosure in central Ethiopia, offers an opportunity to study how different dispersal methods shape forest structure and inform conservation efforts. This study looked at the types of woody species present, their population structure, regeneration status, and how they are dispersed to better understand what supports forest sustainability.

Methods

Woody species were surveyed along seven transects, using 28 plots measuring 20 × 20 m, with nested 2 × 2 m subplots. GPS randomization made sure the transects were placed without bias. Plants were grouped as seedlings, saplings, or mature based on their height and diameter. Importance Value Index (IVI), Shannon diversity, basal area dominance (RDO), regeneration ratios, and species dominance were computed. Dispersal syndromes were assigned using fruit and seed features and regional plant guides. Statistical analyses were performed using ANOVA, Kruskal–Wallis, correlations, regression, and PCA to explore links between dispersal types, IVI, and regeneration.

Results

27 woody species from 15 families were recorded, with Fabaceae dominating. Species diversity was moderate (H′ = 2.84). Endozoochory was the dominant dispersal mode (48.15%), followed by anemochory (22.22%) and autochory (18.52%). Species density strongly predicted IVI (r = 0.92), and PCA revealed that endozoochorous species clustered with high ecological importance. Basal area dominance and IVI-based clustering identified four species (e.g., Ficus sycomorus, Vachellia sieberiana) with regeneration failure, while pioneers such as Dodonaea angustifolia showed strong recruitment. Overall, Biyo Forest recorded 1288.4 seedlings/ha, 836.6 saplings/ha, and 1013.4 mature individuals.

Conclusion

Biyo Forest is dominated by animal-dispersed species, emphasizing frugivores’ role in forest dynamics. Several species face regeneration bottlenecks, indicating the need for targeted conservation, protection of keystone species, and maintenance of frugivore networks to sustain ecological resilience.

Clinical trial number

Not applicable.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12862-026-02580-2.

Introduction

Seed dispersal is essential for the regeneration of forests, as it affects not only where seeds arrive but also the likelihood that seedlings will establish, survive, and contribute to shaping future forests. Therefore, dispersal ecology provides a mechanistic link among plant reproductive strategies, community assembly, species coexistence, and ecosystem resilience. In tropical forests, variations in dispersal mechanisms affect recruitment success, gene flow, and recovery after disturbance by controlling which species may reach appropriate microhabitats [1–3].

Understanding dispersal ecology has become increasingly important for projecting how vegetation will respond to habitat degradation, fragmentation, biological invasions, and climate change since dispersal affects functional diversity as well as population processes [4]. Dispersal syndromes, endzoochory, anemochory, autochory, barochory, hydrochory, and myrmecochory are evolutionary adaptations that affect recruitment success, population dynamics, and spatial patterns of seed movement [1–3, 5]. Consequently, integrating dispersal ecology with regeneration assessments provides a more mechanistic understanding of forest resilience than floristic inventories alone.

Tropical dry forests and woodlands best show the ecological significance of dispersal, as moisture stress, grazing pressure, repeated fire, and human disturbances usually limit natural regeneration [6, 7]. Successful forest recovery in such ecosystems relies not just on the preservation of the current flora but also on the ongoing influx of viable seeds from nearby scenes. Many tropical woody species depend on frugivorous birds, bats, primates, and other animals to carry seeds over great distances, so preserving gene flow, improving colonization of disturbed areas, and lowering density-dependent death under parent trees depend on animal-mediated seed dispersal [8]. However, habitat loss, defaunation, agricultural development, and overexploitation increasingly jeopardize these disperser populations [9]. Declines in seed dispersers can compromise ecosystem resilience, change species composition, lower functional diversity, and thereby impede recruitment. Therefore, knowing the reliance of woody species on several dispersal mechanisms helps to forecast restoration results, find regeneration bottlenecks, and create conservation plans therefore safeguarding plant communities and their ecological relationships.

Dry Afromontane forests and woodlands constitute one of Ethiopia’s most threatened vegetation types while simultaneously providing critical ecosystem services and supporting rural livelihoods [10]. These forests occupy extensive portions of the country’s semi-arid and sub-humid landscapes [7] and contribute substantially to biodiversity conservation, carbon sequestration, soil stabilization, hydrological regulation, and the provision of timber, fuelwood, medicinal plants, and non-timber forest products [10]. They also serve as important refugia for diverse woody plant species, many of which possess specialized dispersal adaptations that enable persistence under harsh environmental conditions and recurrent disturbances [11, 12]. Despite their ecological importance, these forests continue to decline because of agricultural expansion, fuelwood harvesting, overgrazing, settlement encroachment, and infrastructure development [13, 14]. The resulting fragmentation reduces habitat connectivity, limits seed dispersal, and threatens the ecological processes necessary for long-term forest regeneration. To counteract these pressures, Ethiopia has invested extensively in area exclosures as one of its principal landscape restoration strategies over the past three decades. Exclosures have been widely implemented throughout northern and central Ethiopia to facilitate natural regeneration by excluding livestock grazing and minimizing human disturbance, resulting in improved vegetation cover, enhanced soil quality, increased carbon storage, and greater woody species recruitment [15–17]. Nevertheless, restoration success depends not only on disturbance exclusion but also on the ecological processes that enable species recolonization. Seed arrival, dispersal distance, and successful establishment determine whether degraded sites recover their original species composition or become dominated by a limited number of pioneer taxa [18, 19]. Species with effective long-distance dispersal may recolonize restored habitats rapidly, whereas those with restricted dispersal may remain absent despite improved environmental conditions [20, 21]. Consequently, understanding dispersal ecology is essential for evaluating restoration effectiveness, predicting vegetation trajectories, and developing scientifically informed forest restoration programmes. Although numerous studies in Ethiopia have documented floristic composition, woody species diversity, stand structure, and regeneration status across different forest ecosystems [22–25], relatively few have explicitly examined how dispersal syndromes influence species composition, regeneration dynamics, and ecological dominance [26, 27]. Existing studies have largely emphasized structural attributes of vegetation without considering the ecological processes that determine recruitment and population persistence. Consequently, important questions remain regarding how different dispersal strategies influence regeneration success, species dominance, and long-term resilience of dry Afromontane forests. Furthermore, while studies from other tropical regions have demonstrated that disruption of seed dispersal networks can fundamentally alter forest composition and restoration outcomes, comparable evidence from Ethiopian dry forests remains scarce. Addressing this knowledge gap is essential for understanding the mechanisms underlying vegetation recovery and for developing conservation strategies that maintain both plant diversity and the ecological functions supporting forest resilience. Biyo Forest, a remnant dry Afromontane woodland protected through community-based exclosure management [28], provides an ideal natural laboratory for investigating these relationships. Although relatively small in extent, the forest functions as an important biodiversity refuge and seed source within an increasingly fragmented landscape. Its combination of long-term protection, ongoing natural regeneration, and minimal scientific documentation offers a unique opportunity to examine how dispersal mechanisms influence woody species composition, population structure, and regeneration dynamics within a restored dry forest ecosystem. Such information is critical for identifying species experiencing regeneration limitations, understanding the ecological importance of different dispersal strategies, and guiding restoration interventions tailored to species-specific ecological requirements.

Therefore, this study aimed to: (1) characterize the woody species composition and population structure of Biyo Forest; (2) evaluate the regeneration status of its woody species across size classes; and (3) analyze the distribution of dispersal syndromes and their relationships with species density and regeneration dynamics. By integrating vegetation structure, regeneration ecology, and dispersal biology, this study provides new insights into the ecological processes governing forest recovery and offers an evidence-based framework for biodiversity conservation, restoration planning, and sustainable management of dry Afromontane forests in Ethiopia and other tropical dry forest ecosystems.

Materials and methods

Description of the study area

The study was carried out in Biyo forest, a community-managed dry Afromontane forest located in Lumie district of East Shewa zone, Oromia National Regional State, central Ethiopia, (8⁰31’30^’’- 8^o42’0’’N and 38⁰31’30^’’ -39^o 42’0’’E) approximately 62 km east of Addis Ababa (Fig. 1). The district covers approximately 75,220 ha, of which about 1,462 ha is forested, while Biyo Forest itself occupies approximately 70 ha.

Fig. 1.

Map of the study area

The vegetation types of the study area have been categorized under the ‘‘Semi-arid woodland with broad-leaved species’’ which are characterized by several species of Vachellia (V. abyssinica, V. albida, V. seyal, V. tortilis, etc.), Combretum molle, Croton macrostachyus, Dodonaea angustifolia, Erythrina abyssinica, Euphorbia candelabrum, and Olea europaea subsp. cuspidata [29].

Sampling design

Field sampling was conducted during the main growing season from 2 to 25 October 2023, when seedlings, saplings, reproductive structures, and vegetative characteristics could be identified with minimal seasonal bias.

The survey followed a systematic transect-based sampling design widely used in tropical forest inventories. Seven transects were established across the forest following the principal environmental gradient from the foot slope to the summit, covering elevations between 1,876 and 1,960 m a.s.l. Transects were spaced approximately 200 m apart to minimize spatial autocorrelation while maximizing representation of environmental heterogeneity.

To reduce observer bias during transect establishment, the starting point of each transect was determined using GPS-based randomization in QGIS (version 3.34), where random geographic coordinates were generated using the Random Points algorithm. These coordinates served as unbiased starting locations from which transects were established in the field. Transect lengths ranged from 900 m to 1.2 km, depending on topography and accessibility.

Along each transect, twenty-eight sample plots measuring 20 × 20 m (400 m²) were established at 100 m intervals [30] except when encountering a road, yielding a total sampled area of 1.12 ha, representing approximately 1.6% of the forest area.

The selected plot size follows internationally accepted standards for woody vegetation inventories in tropical forests because it provides sufficient representation of tree diversity, stand structure, and species composition while remaining operationally efficient for detailed ecological measurements. Within every main plot, five nested regeneration subplots (2 × 2 m) were established at the four corners and the centre to improve sampling efficiency for seedlings and saplings.

The systematic distribution of plots across the environmental gradient ensured representative sampling of vegetation heterogeneity throughout the forest [31].

Data collection

Within each 20 × 20 m plot, all woody plant species were identified and recorded. For every woody individual having a diameter at breast height (DBH) ≥ 2.5 cm, DBH was measured at 1.3 m above ground level using a diameter tape following standard forest mensuration procedures.

Tree height was measured using a Suunto clinometer for taller trees, whereas individuals shorter than approximately 3 m were measured directly using a graduated measuring pole.

The geographic coordinates and elevation of every sampling plot were recorded using a Garmin eTrex GPS receiver.

Plant identification was carried out using the Flora of Ethiopia and Eritrea and other standard taxonomic references. Whenever field identification was uncertain, voucher specimens were collected, pressed, labelled, and later verified by taxonomic specialists. Voucher specimens (BIF01–BIF27) were deposited in the Teaching Herbarium of Adama Science and Technology University.

Regeneration assessment

Regeneration status was evaluated using height-based developmental classes widely applied in tropical forest ecology. Woody individuals were classified as: Seedlings: height < 1.5 m, Saplings: height 1.5–3.0 m, and Adult trees: height > 3.0 m with DBH ≥ 2.5 cm. These classifications were applied consistently throughout all regeneration analyses.

Species-specific regeneration status was evaluated by comparing the density of seedlings, saplings, and mature individuals. Regeneration ratios (seedlings: saplings: adults) were calculated for each species following [32, 33]. Species exhibiting greater seedling abundance than mature individuals were considered to have good regeneration, whereas species lacking seedlings and/or saplings were interpreted as exhibiting regeneration failure or recruitment limitation.

Assignment of dispersal syndromes

Dispersal syndromes were assigned to each woody species in Biyo Forest using fruit and seed morphology, supported by regional floras, ecological checklists, and studies from Ethiopian drylands. Classification followed standard trait-based approaches.

Species with fleshy, animal-attractive fruits were categorized as endozoochorous; small or winged seeds as anemochorous; large, heavy seeds as barochorous; buoyant seeds as hydrochorous; and seeds with fleshy appendages (elaiosomes) as myrmecochorous [34]. Where multiple dispersal modes were possible, the primary functional syndrome considered most influential for seed movement and regeneration was selected.

Dispersal syndromes were assigned independently by the authors using a trait-based approach based on fruit morphology, seed characteristics, dispersal structures, ecological traits, and information obtained from the Flora of Ethiopia and Eritrea, regional floras, and published ecological literature.

Species were classified into five primary dispersal syndromes: Endozoochory (animal ingestion of fleshy fruits), Anemochory (wind dispersal), Autochory (explosive fruit dehiscence), Barochory (gravity dispersal), and Hydrochory (water dispersal).

When published literature indicated multiple legitimate dispersal mechanisms for a species, only the primary functional dispersal syndrome responsible for long-distance recruitment and population persistence was retained for statistical analyses to ensure classification consistency. Secondary dispersal mechanisms were considered qualitatively during ecological interpretation but were not included in quantitative analyses.

Data analysis

Structural analysis

Basal area of each woody species was calculated as:

Where: BA=Basal Area in m² per hectare d=diameter at breast height (m)

π = 3.14

Species density (stems ha⁻¹), frequency, relative density (RD), relative frequency (RF), and relative dominance (RDO) were calculated. The ecological importance of each species was assessed using the Importance Value Index (IVI):

[35, 36], IVI = RD+RDO + RF

Species were subsequently classified into ecological dominance categories based on IVI values: Dominant (IVI > 20), Common (10–20), Rare (< 10), following commonly adopted thresholds in Ethiopian forest ecology [27, 37].

Diversity and regeneration analyses

Species diversity was quantified using the Shannon–Wiener diversity index (H′) together with Pielou’s evenness index (J). Floristic similarity among plots was assessed using similarity coefficients [31]. Forest regeneration status was evaluated from species-specific seedling, sapling, and adult densities together with regeneration ratios (seedling/sapling; seedling/adult; sapling/adult).

Statistical analysis

All statistical analyses were performed in R version 4.3.2 [38].

Prior to hypothesis testing, assumptions of normality and homogeneity of variance were examined using the Shapiro–Wilk and Levene’s tests, respectively.

To evaluate whether the sampling intensity adequately represented woody species diversity, a species accumulation (species-area) curve was generated using the vegan package in R. The curve approached an asymptote, indicating that additional plots would likely contribute relatively few additional species and supporting the adequacy of the sampling design (Figure S1).

Differences in Importance Value Index among dispersal syndromes were evaluated using one-way analysis of variance (ANOVA), considering species as the unit of analysis. Each woody species represented one independent observation and was grouped according to its primary dispersal syndrome. Only dispersal categories represented by at least three species were included in ANOVA. When ANOVA assumptions were violated, the Kruskal–Wallis test was applied. Significant differences among groups were identified using Tukey’s Honestly Significant Difference (HSD) or Dunn’s multiple comparison test, respectively.

Relationships among IVI, relative density, relative frequency, and relative dominance were examined using Pearson or Spearman correlation coefficients, depending on data normality. Multiple linear regression was subsequently performed to determine the relative contribution of RD, RF, and RDO to IVI. Multicollinearity among explanatory variables was evaluated using Variance Inflation Factors (VIF).

To investigate multivariate relationships among species, Principal Component Analysis (PCA) was conducted using standardized (z-score transformed) RD, RF, and RDO values. Variables were centered and scaled to unit variance prior to analysis to eliminate differences in measurement scales. Species scores were projected onto the first two principal components to visualize relationships among species, structural attributes, and dispersal syndromes. Figures were produced using the ggplot2 package, while diversity indices were calculated using the vegan package.

Results

Species composition and diversity

A total of 27 woody species belonging to 23 genera and 15 families were recorded in Biyo Forest. Fabaceae was the most species-rich family, contributing ten species (37.0%), followed by Euphorbiaceae (three species) and Lamiaceae (two species), while the remaining families were each represented by a single species (Table S1).

Assessment of the recorded woody species against the IUCN Red List showed that 21 species (77.8%) are classified as Least Concern (LC), while six species (22.2%) have Not Evaluated (NE) status. None of the species recorded in the study area are currently listed as globally threatened (Vulnerable, Endangered, or Critically Endangered). Furthermore, none of the identified species are endemic to Ethiopia. Most species are native and widely distributed across eastern and sub-Saharan Africa, whereas Acacia saligna, Acacia melanoxylon, and Schinus molle are introduced species in Ethiopia. Although no Ethiopian endemic species were recorded, several taxa such as Grewia ferruginea, Rhus retinorrhoea, and Vernonia ischnophylla are characteristic of the Horn of Africa flora and contribute substantially to the ecological uniqueness of the region (Table S1). The Shannon–Wiener diversity index was 2.84, with an evenness value of 0.63, indicating moderate woody species diversity and a relatively balanced distribution of individuals among species.

Forest structural characteristics and species dominance

Importance Value Index (IVI) analysis showed that forest structure was dominated by relatively few species. Vachellia tortilis had the highest IVI (46.01), followed by Dodonaea angustifolia (27.73), Grewia ferruginea (25.94), Vachellia abyssinica (22.69), and Vachellia etbaica (20.43). Together, these five species accounted for approximately 47.6% of the total IVI (Table 1), whereas the remaining 22 species collectively contributed 52.4%.

Table 1.

Importance Value Index (IVI) of the dominant tree species of Biyo forest (RF = Relative Frequency, RD = Relative Density, and RDO = Relative Dominance IVI = Important Value Index) arranged in descending order of their IVI Values

Species	RD	RF	RDO	IVI
Vachellia tortilis	8.41	9.44	28.16	46.01
Dodonaea angustifolia	17.98	6.67	3.08	27.73
Grewia ferruginea	12.43	9.44	4.07	25.94
Vachellia abyssinica	4.17	4.44	14.08	22.69
Vachellia etbaica	5.41	6.67	8.35	20.43
Rhus retinorrhoea	6.29	6.11	3.35	15.75
Acacia saligna	6.65	5	1.1	12.75
Acacia melanoxylon	3.14	5.56	3.2	11.9
Ficus sycomorus	0.07	0.56	10.85	11.48
Vachellia sieberiana	3.8	6.11	1.55	11.46
Vachellia seyal	3.07	2.78	5.17	11.02
Aeschynomene elaphroxylon	3.14	4.44	2.63	10.21
Combretum collinum	1.75	3.33	3.39	8.47
Calpurina aurea	4.17	2.78	1.51	8.46
Ehreta cymosa	2.85	2.78	1.65	7.28
Capparis tomentosa	1.91	2.78	1.95	6.64
Croton macrostachyus	1.02	3.88	1.52	6.42
Clerodendrum myricoides	2.71	2.78	0.3	5.79
Pterolobium stellatum	2.41	2.22	0.41	5.04
Euphorbia candelabrum	1.02	2.22	1.72	4.96
Ocimum urticifolium	1.39	3.33	0.19	4.91
Euphorbia tirucalli	2.63	1.67	0.26	4.56
Schinus molle	0.8	1.67	0.81	3.28
Justicia schimperiana	1.24	1.11	0.08	2.43
Carissa spinarum	1.1	0.56	0.21	1.87
Vernonia ischnophylla	0.29	1.11	0.39	1.79
Solanum campylacanthum	0.15	0.56	0.02	0.73

Basal-area analysis demonstrated marked variation among species. V. tortilis contributed the largest proportion of stand basal area (RDO = 28.16%), followed by V. abyssinica (14.08%) and Ficus sycomorus (10.85%). Most remaining species exhibited RDO values below 5%, indicating comparatively small contributions to canopy structure (Fig. 2).

Fig. 2.

Basal area dominance curve of woody species in Biyo Forest, Central Ethiopia

Based on IVI thresholds, five species were classified as dominant (IVI > 20), seven as common (IVI = 10–20), and fifteen as rare (IVI < 10), representing approximately 55.6% of all recorded woody species (Fig. 3).

Fig. 3.

A bar plot showing the Importance Value Index (IVI) based species clustering of Biyo Forest

Relationships among structural attributes

Pearson correlation analysis revealed significant positive relationships between IVI and all three structural variables. Relative density showed the strongest correlation with IVI (r = 0.92, p < 0.001), followed by relative frequency (r = 0.74, p = 0.001) and relative dominance (r = 0.58, p = 0.01).

Multiple linear regression explained 89% of the variation in IVI (Adjusted R² =0.89, p < 0.001). Relative density had the largest standardized coefficient (β = 0.73), followed by relative frequency (β = 0.41) and relative dominance (β = 0.28), with variance inflation factors below three, indicating negligible multicollinearity (Fig. 4).

Fig. 4.

Multiple linear regression using IVI as the response variable and RD (a), RDO (b), and RF (c) as predictors

Principal component analysis

Principal Component Analysis based on standardized relative density, frequency, and dominance explained 82.4% of total variation, with PC1 accounting for 61.2% and PC2 for 21.2%.

PC1 was primarily associated with IVI and relative density, whereas PC2 represented variation in relative frequency and dominance. Species dispersed by animals generally occupied the positive side of PC1 and were associated with higher IVI values, whereas wind- and gravity-dispersed species were distributed toward lower PC1 scores (Fig. 5).

Fig. 5.

PCA biplot showing species ordination based on RD, RF, and RDO. Dispersal syndromes are color-coded; where RD=Relative Density, RF= Relative Frequency, and RDO= Relative Dominance

Regeneration status

Among the 18 woody species evaluated for regeneration, total densities were 1288.39 seedlings ha⁻¹, 836.61 saplings ha⁻¹, and 1013.39 mature individuals ha⁻¹. The overall regeneration ratios were 1.27 seedlings and 0.83 saplings per mature individual.

Regeneration differed considerably among species. Four species (Vachellia sieberiana, Croton macrostachyus, Ehretia cymosa, and Ficus sycomorus) lacked both seedlings and saplings, whereas three species (V. abyssinica, V. etbaica, and Combretum collinum) were represented by seedlings but lacked saplings. Dodonaea angustifolia, Grewia ferruginea, and V. tortilis exhibited the highest densities of regenerating individuals (Fig. 6).

Fig. 6.

Density percentage of mature trees, saplings, and seedlings in Biyo Forest

Dispersal syndromes

Five dispersal syndromes were identified among the 27 woody species. Endozoochory was the most common syndrome (48.15%), followed by anemochory (22.22%), autochory (18.52%), barochory (7.41%), and hydrochory (3.70%) (Table 2).

Table 2.

Dispersal syndromes of woody species in Biyo Forest, Lumie District, central Ethiopia, including primary sources used for the assignment based on fruit/seed morphology and ecological observations

Scientific name	Dispersal syndrome	Main dispersal agents	References
Vachellia abyssinica (A. Rich.) Vatke	Endozoochory	Animal ingestion; occasional epizoochory	[39–43]
Vachellia etbaica Schweinf.	Endozoochory	Herbivores, browsers	[39–43]
Acacia melanoxylon R. Br.	Anemochory	Wind	[39, 40, 43]
Acacia saligna (Labill.) Wendl.	Myrmecochory	Ant dispersal via elaiosomes; gravity	[39, 40, 43]
Vachellia seyal Del.	Endozoochory	Ungulates, browsers	[39, 41–47]
Vachellia sieberiana DC.	Endozoochory	Large herbivores	[39, 41–43]
Vachellia tortilis (Forssk.) Hayne	Endozoochory	Large herbivores	[39, 41–47]
Aeschynomene elaphroxylon (Guill. & Perr.) Taub.	Hydrochory	Water	[39, 40]
Calpurnia aurea (Ait.) Benth.	Autochory	Explosive pod dehiscence; gravity	[39, 40, 41, 47, 49]
Capparis tomentosa Lam.	Endozoochory	Frugivorous birds and mammals	[40, 51]
Carissa spinarum L.	Endozoochory	Birds; fleshy fruits	[39, 40, 52]
Clerodendrum myricoides (Hochst.) Vatke	Endozoochory	Birds consuming drupes	[39, 40, 54]
Combretum collinum Fresen.	Anemochory	Winged samaras dispersed by the wind	[40, 43, 48, 51, 54–56]
Croton macrostachyus Del.	Autochory	Explosive capsule dehiscence	[43, 44, 51, 54, 57]
Dodonaea angustifolia L.f.	Anemochory	Winged seeds	[40, 42, 53, 57]
Ehretia cymosa Thonn.	Endozoochory	Birds; fleshy drupes	[40, 49, 51, 54, 57]
Euphorbia candelabrum Kotschy	Autochory	Explosive capsules; gravity	[40, 51, 57]
Euphorbia tirucalli L.	Autochory	Explosive capsules; gravity	[40, 51, 57]
Ficus sycomorus L.	Endozoochory	Birds, bats, monkeys	[43, 44, 48, 51, 54, 58]
Grewia ferruginea Hochst. ex A. Rich.	Endozoochory	Birds; fleshy drupes	[42, 44, 47, 48, 56]
Justicia schimperiana (Hochst. ex Nees) T. Anders.	Autochory /Myrmecochory	Explosive capsules; occasional ant dispersal	[39, 40, 48, 49, 54]
Ocimum urticifolium	Barochory	Gravity; small nutlets	[40, 50]
Pterolobium stellatum	Anemochory	Winged samaroid pods	[39, 40, 42]
Rhus retinorrhoea	Barochory	Gravity with opportunistic removal	[39, 44, 54]
Schinus molle	Endozoochory	Vertebrates	[39, 40]
Solanum campylacanthum	Endozoochory	Mammals; large herbivores	[53, 59]
Vernonia ischnophylla	Anemochory	Wind-dispersed pappus achenes	[40, 45, 60]

Relationships between dispersal syndromes, ecological importance and regeneration

IVI differed significantly among dispersal syndromes (ANOVA: F = 8.72, p = 0.002). Endozoochorous species exhibited significantly greater IVI than barochorous species, while differences between endozoochorous and anemochorous species were marginally significant (Fig. 7).

Fig. 7.

Boxplot illustrating the distribution of Importance Value Index (IVI) across dispersal syndromes (endozoochory, anemochory, Autochory), for syndromes represented by at least three species

Recruitment also differed among dispersal syndromes. Endozoochorous species generally recorded higher seedling and sapling densities than other dispersal groups, whereas autochorous, barochorous, and hydrochorous species exhibited comparatively lower recruitment. ANOVA showed significant effects of dispersal syndrome on both seedling density (F₄,₂₀ =17.41, p < 0.001) and sapling density (F₄,₂₀ =32.63, p < 0.001, Fig. 8).

Fig. 8.

Effects of dispersal mode on the regeneration of woody species

Discussion

Species composition, forest structure and ecological significance

The woody species composition and the structural organization of Biyo Forest represented a community shaped by species-specific functional traits, ecological adaptation, and the long history of anthropogenic disturbance. Although the woody flora was moderately diverse, the structure of Biyo Forest was disproportionately influenced by the relatively few species with high Importance Value Indices (IVI). This pattern implies that community organization is governed more by functional dominance than species richness, a common feature of seasonally dry tropical forests, where environmental filtering and recurrent disturbance favor a small number of ecologically competitive species. Similar dominance patterns are reported from dry Afromontane forests elsewhere in Ethiopia, where a small number of drought-tolerant woody species comprise much of the stand biomass and canopy structure [61, 62].

Fabaceae was the dominant family, with ten woody species in different ecological niches and with different dispersal syndromes. This family is dominant due to its wide ecological amplitude and adaptive features such as nitrogen fixation, drought tolerance, resource acquisition efficiency, and resilience to repeated disturbance. Seed dispersal mediated by vertebrates also benefits several species of Fabaceae, favoring colonization and persistence in heterogeneous environments. Similar patterns have been reported from Belete Forest [61], Hugumburda Forest [62], and Komto Forest [11], where Fabaceae has always been the structural backbone of Ethiopian Afromontane Forest ecosystems. A small number of species dominated the forest, in particular Vachellia tortilis, Dodonaea angustifolia, Grewia ferruginea, Vachellia abyssinica and Vachellia etbaica, which together accounted for almost half the total IVI. The very high IVI and basal area of V. tortilis indicate better competitive ability under water-limited conditions. V. abyssinica, on the other hand, contributes significantly to the canopy structure though it has relatively lower stem density. Such skewed dominance distributions are typical of dry woodlands where a few species maintain structural dominance through drought tolerance, the ability to coppice, and efficient resource acquisition, despite environmental stress [25, 63]. Ethiopian dry woodlands show similar ecological patterns with Vachellia species consistently dominating due to their ecological plasticity, resistance to disturbance, and high functional value to surrounding communities [64, 65].

An interesting contrast was observed for Ficus sycomorus, where the species had relatively low density and frequency but contributed disproportionately to stand basal area. This pattern reflects the ecological importance of large fig trees as keystone species, with a far greater impact than their numbers would suggest. Mature trees provide fruit year-round to frugivorous birds, bats and mammals, facilitate long-distance seed dispersal, provide favorable microhabitats for seedling establishment and are important contributors to forest carbon storage [58, 66–68]. Similar observations from the Bale Mountains [69] and Tara Gedam Forest [70] suggest that the removal of large fig trees may have cascading effects on frugivore communities, reduce seed dispersal services and ultimately compromise regeneration of numerous fleshy-fruited woody species.

Dispersal strategy as determinant of ecological success

One of the main outcomes of this study is that the seed dispersal strategy strongly affects ecological importance and regeneration success. Species dispersed by vertebrates (endozoochory) were in the highest IVI classes, grouped together in the multivariate analyses, and generally had stronger regeneration than species using other dispersal mechanisms. The results suggest that dispersal strategy affects long-term community assembly through its influences on recruitment, spatial distribution, competitive ability and persistence of populations, as well as seed movement. Endozoochory was the most dominant mode of dispersal (48.15% of all recorded woody species), and this emphasizes the importance of frugivorous birds, bats and medium-sized mammals in the maintenance of forest regeneration in Biyo Forest. Animal-mediated seed dispersal contributes to long-distance movement of seeds, decreases density-dependent mortality under parent trees, improves genetic connectivity and facilitates colonization of favourable regeneration microsites. Similar patterns of associations between vertebrate dispersers and recruitment of woody plants have been reported for Ethiopian Afromontane forests [71, 72], Kakamega Forest in Kenya [73] and Kibale National Park in Uganda [74]. In these systems, disperser communities are key to the maintenance of forest diversity in the face of recurrent disturbance.

The positive association between endozoochory and high IVI further suggests that effective dispersal enhances ecological dominance through improved recruitment and population persistence. Species such as Grewia ferruginea and Vachellia tortilis combined high regeneration success with high ecological importance, demonstrating how efficient seed dispersal contributes to both stand structure and long-term population stability. Similar relationships between dispersal effectiveness and forest composition have been reported from southwestern Ethiopia [75] and Kakamega Forest [76], emphasizing the importance of vertebrate-mediated dispersal in tropical forest dynamics.

Regeneration is regulated by both dispersal and post-dispersal ecological processes

Though most endozoochorous species showed more regeneration, Biyo Forest’s regeneration patterns show that effective recruitment is not always dependent on good seed dispersion alone. The example of Ficus sycomorus perfectly shows this ecological idea. Though commonly spread by frugivorous creatures and major structural species in the woods, it was missing from the tested sites seedlings and saplings. This obvious regeneration bottleneck implies that post-dispersal ecological filters, including microsite suitability, soil moisture availability, seed predation, herbivory, pathogen pressure, and species-specific germination requirements, also affect recruitment in addition to seed arrival. Emphasizing that several ecological processes influence successful forest regeneration after seed dispersal has taken place, similar recruiting restrictions have been noted for large-seeded tropical tree species in African and Asian woodlands [58, 68, 72, 76, 77].

Species dispersed by wind showed different regeneration techniques. For instance, Dodonaea angustifolia generates many seedlings and high stem densities but only little to total basal area, therefore reflecting its function as a pioneer species able to quickly colonize disturbed areas and forest gaps. Such species dominate early successional phases, but they are progressively superseded as slower-growing canopy species take hold. Similar successional processes have been observed in Bale Mountains [69], Tara Gedam Forest [70], and Wof Washa Forest [78], where Dodonaea angustifolia always defines secondary succession after disturbance.

Conversely, autochorous and barochorous species showed persistently poor regeneration and somewhat low seedling and sapling numbers. Their limited dispersal distances focus young near parent trees, therefore raising vulnerability to density-dependent death, herbivory, disease assault, and resource scarcity competition. These restrictions may be made worse by livestock grazing, disruption of undergrowth, and loss of habitat seen in the study area. Bengi Forest [79], Menagesha Amba Mariam Forest [71], Kibale Forest [74], and Amazonian tropical forests [80] are said to have comparable regeneration limits. Similarly, the only hydrochorous species (Aeschynomene elaphroxylon) showed little recruitment, therefore highlighting the somewhat restricted ecological significance of hydrochory in upland dry forest environments [81].

Implications for restoration and conservation

The patterns of regeneration seen in Biyo Forest suggest that the long-term resilience of a forest depends on both maintaining woody species diversity and conserving the ecological processes that support regeneration. Particularly, the predominance and ecological success of endozoochorous species highlight the need for robust populations of frugivorous birds, bats, and mammals to preserve working seed dispersal networks and help to naturally restore forests. Therefore, restoration initiatives that concentrate only on tree planting without safeguarding disperser populations could neglect to recover the long-term environmental resilience-required interactions.

The poor regeneration seen in some economically significant species also points to the possibility that passive protection by itself might not be enough to preserve Biyo Forest’s whole diversity. Improving recruitment and long-term population viability would probably result from strengthening community-managed exclosures, lowering animal grazing, preserving mature seed trees, applying assisted natural regeneration, and enhancing numbers of poorly regenerating species. Keystone species such as Ficus sycomorus, whose ecological roles go beyond structural dominance by supporting frugivore numbers, enabling long-distance seed dispersion, and preserving wider ecosystem processes [58, 66–68], deserve special conservation attention.

From a conservation perspective, the IUCN Red List assessment shows that most of the woody species found in the study area are classified as Least Concern (LC), therefore indicating that they are not now thought to be at great risk of global extinction. However, under the IUCN Red List guidelines, a number of native species such as Grewia ferruginea, Rhus retinorrhoea, Clerodendrum myricoides, Ocimum urticifolium, Justicia schimperiana, and Vernonia ischnophylla remain Not Evaluated (NE). Furthermore, none of the documented species are presently ranked as Near Threatened, Vulnerable, Endangered, or Critically Endangered, and no Ethiopian endemic woody species were noted in the study area. Though these results point to a rather good worldwide conservation status, they should not be taken as proof that local populations are safe. Global evaluations might miss regional losses brought on by habitat fragmentation, degradation, overexploitation, and lower regeneration.

The ecological importance of locally dominant species further reinforces this point. Species with high Importance Value Index (IVI), particularly Vachellia tortilis, Dodonaea angustifolia, and Grewia ferruginea, make substantial contributions to forest structure, ecosystem functioning, and habitat quality. Their persistence is therefore critical for maintaining ecosystem stability and biodiversity. Conservation planning should integrate local ecological indicators, such as regeneration status and IVI, with global conservation assessments to identify species requiring priority management. Such an integrated approach will enable restoration efforts to focus not only on globally threatened taxa but also on ecologically important species whose local populations may be declining despite their favorable IUCN status. This strategy is particularly relevant for dry Afromontane forests, where maintaining ecological integrity depends on preserving species interactions, ensuring continuous natural regeneration, and sustaining the ecosystem functions that underpin long-term forest resilience.

Conclusion

This study achieved three objectives. First, it characterized the woody species composition and structural organization of Biyo Forest, demonstrating moderate diversity but strong dominance by a few ecologically important species. Second, it evaluated regeneration across woody species and identified substantial variation in recruitment, including bottleneck regeneration in several structurally important taxa. Third, it showed that dispersal syndromes, particularly endozoochory, are strongly associated with ecological importance and regeneration success. Collectively, these findings demonstrate that dispersal ecology provides valuable insight into the mechanisms governing forest recovery and should be incorporated into restoration planning for dry Afromontane forests. Although dispersal syndromes were inferred from species traits rather than directly observed and the study was confined to a single forest, the results provide an important baseline for future investigations integrating plant–animal interactions, disperser communities, and long-term regeneration monitoring.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

All authors contributed to the study conception and design. Lemessa Kumsa, and Habte Telila, performed material preparation, data collection and analysis. Lemessa Kumsa wrote the first draft of the manuscript and all authors commented on previous versions of the manuscript. All authors read and approved the final manuscript.

Funding

The authors declare that no funds were received for this study.

Data availability

The data sets supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

Ethical clearance for this study was granted by the Research and Ethics Committee of Adama Science and Technology University (Approval No. ASTU 7841/23). Permission to conduct field surveys and collect plant material was obtained from the relevant regional authorities and local landowners in the fragmented coffee landscapes of Southwest Ethiopia. All plant materials used in this study were collected from wild populations within a community-managed forest. The research adhered to all institutional, national, and international guidelines for the study of plants, including the IUCN Policy Statement on Research Involving Species at Risk of Extinction and Ethiopian biodiversity and access regulations. Dr. Lamessa Kumsa and Dr. Habte Telila, who co-authored this paper, identified the plant species. We placed voucher specimens in the Teaching Herbarium at Adama Science and Technology University with accession numbers BIF01 to BIF27.

Consent for publication

Not applicable. No individual data requiring consent for publication were collected in this study.

Competing interests

The authors declare no competing interests.