Woody Species Diversity, Vegetation Structure, and Regeneration Status of the Moist Afromontane Forest of Agama in Southwestern Ethiopia

This study was conducted in Agama Forest in Kafa Zone, Southwestern Ethiopia, to assess species diversity, vegetation structure, and regeneration status of woody species. A systematic sampling technique was employed to collect vegetation data. Sixty (60) sample plots of 25m × 25m were laid at 300m intervals all along ten grids interspaced 800m apart. Sample plots of 25m × 25m were used to record DBH and H of all woody plant species reaching a DBH > 2.5cm and height > 2m. For the inventory of seedling and sapling, two subplots of 2m × 5m were used at the beginning and the end of the baseline on opposite sides of the main quadrat. Vegetation data such as DBH, height, seedling, and sapling density of woody species were recorded in each plot. Altogether, 72 woody plant species of 65 genera and 35 families were identiﬁed. Analysis of selected tree species showed diverse population structures. This study showed that small trees and shrubs dominated the Agama Forest, which revealed its status under a secondary regeneration stage. Study on the structure and regeneration of some woody species indicated that there are species that require urgent conservation measures. Sound management and monitoring, as well as maintenance of biodiversity and cultural and economic values of the forest, require conservation activities that encourage sustainable uses of the forest and its products.


Introduction
Ethiopia's highly variable ecology, topography, and climate make it an internationally recognized centre of biodiversity [1]. e country has around 6000 higher plant species of which about 10% are endemic [2,3]. e vegetation of Ethiopia has been classified into 12 types [4]. e vegetation type at Agama Forest in Southwestern Ethiopia, the subject of this paper, is part of the moist evergreen Afromontane forest that is characterized by one or more closed strata of evergreen trees that may reach heights of 30 to 40 m.
Southwestern Ethiopia best represents remnant natural forests but those are being destroyed at an alarming rate [5]. Human-induced loss of forest cover, structure, and biodiversity is of global concern; in Ethiopia, [6] estimated rates of deforestation and forest degradation at between 150,000 and 200,000 ha/year and this was associated with loss of forest structure, diversity, dynamics, and evolution. New investment opportunities in Southwestern Ethiopia are converting these remnant forests into other land uses such as tea and coffee plantations [7]. New settlers migrating from the northern and central parts of Ethiopia have also contributed to land use changes and forest degradation [8,9]. e Shannon-Wiener index, H, is the most popular measure of species diversity because it scores for both species richness and evenness and is not affected by sample size [10,11]. In the analysis of vegetation structure, the growth stages of trees as seedlings, saplings, and mature trees as well as the distribution of size classes within a population can be essential elements of diversity that permit or deny the likelihood of quick recovery after disturbances [12]. e status and dynamics of woody-tree populations can be examined by conducting a size class distribution and seedling and sapling counts [13,14]. Healthy natural populations with continuous regeneration exhibit an exponentially decaying size class distribution, whereby trees in smaller size classes are represented in greater numbers than in larger classes. e absence or rarity of seedlings can be considered an indication of a declining population. e population structure of a tree species is indicative of its history of past disturbance and can be used to predict its future status in the forest [13]. is study investigated the woody species diversity, structure, and regeneration status of the Agama Forest in Southwestern Ethiopia. e results will be used to set conservation and management strategies for this forest.

Study Area.
is study was conducted in Gimbo district of the Kafa Zone in the Southern Nations' Nationalities and Peoples' Regional State (SNNPRS), which is located 500 km from Addis Ababa and 30 km from Bonga ( Figure 1). e area is centered at 7.16°N, 36.11°E, the altitudinal range is from 1800 m to 2370 m, and the topography is undulating, with valleys and rolling plateaus [15]. e size of the study forest covers about 1872 hectares (Figure 1). e climate data between the years 2005 and 2018 recorded by the meteorological station at Bonga that is located 20 km south of the study area was used to describe the climatic condition of the study area. ere is a unimodal rainfall pattern with eight months between March and October with rainfall >100 mm/month [16]. e mean annual rainfall is 1830 mm, and the monthly mean maximum and mean minimum temperatures are 29.6°C and 9.5°C, respectively. e mean annual temperature is 19.7°C. e major soil groups of the study area, according to the FAO/UNESCO legend of soil classification, are Nitisols, Acrisols, and Vertisols [17]. e Nitisols are agriculturally the most important and dominant type of soils in the Kafa Zone.
e Nitisols are clay-red in color and have moderate CEC and relatively high organic matter content and total nitrogen.

Vegetation Sampling.
A preliminary survey was made from 30 April 2017 to 15 May 2017 to obtain an impression on the general physiognomy of the vegetation and identify sampling sites in the study area. e actual field study was conducted from 10 December 2017 to 30 April 2018. e systematic sampling design was used to collect vegetation and environmental data [10,18,19]. Sixty (60) sample quadrats of 25 × 25 m were laid at 300 m intervals along ten grids interspaced 800 m apart. Seedling and sapling inventories of all woody-tree and shrub species were recorded in two 2 × 5 m subquadrats located on opposite sides of each quadrat. For all woody species of height (H) ≥ 2 m and diameter at breast height (DBH) ≥ 2.5 cm, H and DBH were measured using a clinometer and diameter tape, respectively. Regeneration patterns were assessed using the total count of seedlings (H ≤ 50 cm and DBH ≤ 2.5 cm) and saplings (H > 50 cm and DBH ≤ 2.5 cm) within the subquadrats. Geographical coordinates and altitudes were recorded for each quadrat using GPS. Plant specimens were collected, pressed, dried, and brought to the National Herbarium (ETH), Addis Ababa University, for taxonomic identification and nomenclature. ese were determined by comparison with authenticated specimens housed at ETH and by referring to published volumes of the Flora of Ethiopia and Eritrea [20][21][22][23][24][25].
2.3. Data Analysis. Species diversity was calculated using the Shannon-Wiener diversity index, H, as where s is the number of woody species and p i is the proportion of individuals or the abundance of the i th species expressed as a proportion of the total. Shannon's evenness, J, was calculated as the ratio of observed diversity, H, to the maximum diversity, H max , using the following equation: where ln(s) � H max e structure of the vegetation was described using a frequency distribution of H, DBH, and Importance Value Index (IVI). Tree or shrub density and basal area values were calculated on a per-hectare basis. For all species, IVIs were calculated as the sum of their relative density (RD), relative frequency (RF), and relative dominance (RDO) [10] where where the basal area of an individual wasπd 2 /4 (π � 3.14; d � DBH). Five IVI classes were established: I < 1; II � 1-10; III ≥ 10-20; IV ≥ 20-30; V > 30.
Frequency (F) and density (D) were calculated as [26] F � no. of quadrats in which a species occurs total no. of quadrats examined × 100. (4) Species were grouped into five frequency classes: total no. of individuals of a species found total area examined .
To interpret the dynamics of woody species in the forest, the population structures of selected species were expressed as a frequency of individuals against established DBH classes. e emerging patterns of diameter class distribution were also used to interpret the recruitment processes of a given species. Species were divided into seven H and DBH classes. To use the regeneration analysis for priority setting, the species considered in the study area were classified into three groups based on the density of the total regeneration.

Woody Species Diversity.
Seventy-two woody plant species belonging to 35 families and 65 genera were recorded (Table 1). Of these, 43 species were trees, 18 were shrubs, and 12 were lianas. Rubiaceae was the most common family with 8 (11.0% contribution) species in 8 (12.3%) genera. Acanthaceae and Euphorbiaceae were the second most common, each with 5 (6.9%) species in 5 (7.7%) genera. e Rutaceae had 4 species in 4 genera and the Fabaceae 4 species in 3 genera. e Araliaceae, Celastraceae, Dracenaceae, Oleaceae, and Rosaceae contributed 3 species each and the remaining 25 families <3 species each. e Shannon-Wiener diversity index and Shannon's evenness values were 3. 25   International Journal of Ecology e D of trees and shrubs with DBH 10−20 cm and DBH > 20 cm were, respectively, 556 and 281 individuals per hectare. Accordingly, the ratio of individuals with DBH 10−20 cm (a) to DBH > 20 cm (b) was 2.0.
Comparison of trees and shrub densities with DBH 10-20 cm (a), DBH > 20 cm (b), and the ratio (a/b) for Agama Forest with 5 other forests in Ethiopia is given in Table 2

Height and DBH Class Distribution.
e H and DBH class distributions of all individuals in the different size classes were an inverted J shape. us the majority of species had the greatest number of individuals with relatively low H and DBH with a gradual decrease in numbers of both with increasing H and DBH. About 63% of individuals were found in the first height class (2.5-5 m); only a few individuals, about 1%, attained heights > 30 m. DBH distribution showed that about 81% of individuals were in the DBH class <20 cm, and a very small proportion (1.2%) reached DBH > 110 cm. ere were four main patterns of population structure (Figures 2(a)-2(d)).
e first was a bell-shaped distribution, in which the number of individuals in the middle diameter classes is highest, e.g., Olea welwitschii and Syzygium guineense. e second was a J-shaped distribution, in which the number of individuals increases with diameter class, e.g., Schefflera abyssinica, Cyathea manniana, and Rytigynia neglecta. e third was formed with species showing an inverted J shape, a pattern where the highest frequency is in the lower diameter classes and it decreases towards the higher diameter classes, e.g., Elaeodendron Buchananii, Olea capensis, Vepris dainellii, Oxyanthus speciosus, and Teclea nobilis. e fourth pattern had no individual in DBH class one, an abrupt increase from DBH class two to the middle classes, and an abrupt decrease from the middle to the higher classes, e.g., Sapium ellipticum, Croton macrostachyus, and Poliscas fulva (Figure 2(d)).
ree priority classes based on the total density of seedlings and saplings were established for conservation: class 1 with no seedling or sapling, class 2 with density of seedlings and saplings >0 but <50 individuals ha −1 , and class 3 with density of seedlings and saplings >50 individuals ha −1 ( Table 4).

Floristic Diversity.
Seventy-two (72) species of shrubs, lianas, or trees were recorded. Among the tree species, two, Vepris dainellii and Millettia ferruginea, are endemic to Ethiopia. e Shannon-Weiner diversity and evenness indexes were 3.25 and 0.78, respectively. us, the species mix at Agama is representative of a forest with high species diversity [10] and the species are well represented across the extent of the forest.

Vegetation Structure.
e ratio of DBH > 10 cm to DBH > 20 cm (a/b ratio) was 2.0 and indicative of the predominance of small-sized individuals in the forest. is was largely due to the high density of Olea capensis and Vepris dainellii. is ratio can also be used as a measure of size class distribution [33] and shows that, in Ethiopia, Agama Forest is comparable to moist Afromontane forests at Belete, Masha, and Menna Angetu, but with a greater predominance of small-sized individuals than those at Gelesha and Komto. at the proportion of individuals of DBH between 2.5 and ≤10 cm was 42.1% suggests that Agama Forest is in a secondary stage of development.
Frequency provides an approximate indication of the homogeneity of a stand [10]. High values in higher frequency classes (D and E) and low values in lower frequency classes (A and b) indicate constant or similar species composition. On the other hand, high values in lower frequency classes and low values in higher frequency classes indicate a high degree of floristic heterogeneity [27]. In this study, high values were obtained in lower frequency classes, which showed the existence of high degree of floristic heterogeneity in Agama Forest.
Elaeodendron buchananii, Olea welwitschii, Sapium ellipticum, Schefflera abyssinica, and Syzygium guineense occupied >70% of the total basal area and can be considered the most important species in Agama Forest. A basal area provides a better measure of the relative importance of the species than simple stem count [34]. us, species with the largest contribution in a basal area can be considered the most important woody species in the study area.
Important Value Index (IVI) permits a comparison of species in a given forest type and depicts the sociological structure of a population in its totality in the community. It often reflects the extent of dominance, occurrence, and abundance of a given species in relation to other associated species in an area [10]. It is also important to compare the ecological significance of a given species. erefore, it is a good index for summarizing vegetation characteristics and ranking species for management and conservation practices. Important Value Index combines data for three parameters   (relative frequency, relative density, and relative abundance) [35]. We pointed out that Important Value Index gives a more realistic figure of dominance from the structural point of view. It is useful to compare the ecological significance of species [27]. In Agama Forest, distribution of all individuals in different height and DBH classes indicated an inverted J-shaped curve, which shows a normal population structure with a high number of individuals in the lower size classes and only a few individuals in the higher size classes. is pattern is an indicator of healthy regeneration of the forest and species and shows a good reproduction and recruitment capacity. Even though the overall height and DBH distribution revealed inverse J shape, different population dynamics for different species were in this study.
Information on the population structure of a tree species indicates the history of the past disturbance to that species and the environment and, hence, is used to forecast the future trend of the population of that particular species [13]. Population structure is an extremely useful tool for orienting management activities and perhaps most important for assessing both the potential of a given resource and the impacts of resource extraction [13]. In this study, four patterns population distributions based on DBH were revealed for selected woody species. ese are J-shaped, bellshaped, inverted J-shaped, and irregular shaped. e J-shaped patterns show poor reproduction and hampered regeneration due to the fact that either most trees are not producing seeds due to age or there are losses due to predators after reproduction (e.g., Schefflera abyssinica). A bell shape follows a Gauss distribution pattern. is pattern indicates a poor reproduction and recruitment of species, which may be associated with the overharvesting of seed bearing individuals (e.g., Olea welwitschii). Bell-shaped or variable size class distribution has been attributed to a disturbed forest where regeneration is hampered [36]. An inverted J-shaped distribution pattern of species is considered as an indication of stable population status or good regeneration status [37]. An irregular shaped pattern characterized by no individual in DBH class one, with an abrupt increase from DBH class two to the middle classes and with an abrupt decrease from the middle to the higher classes (e.g., Sapium ellipticum), might reflect limited regeneration, possibly due to human disturbance, livestock trampling or browsing, and other biotic and abiotic factors.

Regeneration Status.
Regeneration refers to the process of silvigenesis by which trees and forests survive over time [38]. e population structure along different developmental stage of a species in a forest can express its regeneration behavior [39]. e population structure, described by the existence of sufficient population of seedlings, saplings, and adults, shows successful regeneration of forest species [39].
In this study, five distribution patterns of regeneration status were observed from the 54 woody species investigated for regeneration. (1) Seedling > sapling > tree/shrub state (e.g., Millettia ferruginea): this pattern indicates good regeneration. (2) Seedling > sapling < tree/shrub state (e.g., Elaeodendron buchananii): this pattern represents fair regeneration and recruitment of the species. (3) Seedling < sapling < tree/shrub state (e.g., Ficus sur): this pattern shows poor reproduction and hampered regeneration either due to the fact that most trees are not producing seeds as a result of their old age or there has been a loss of seeds by predators after reproduction. For instance, the fruits of Ficus sur were usually eaten as food by many animals including humans, which might be a reason for this pattern. (4) With no individual either in seedling or sapling stages but relatively many individuals in tree/shrub stage (e.g., Fagaropsis angolensis and Flacourtia indica): this pattern also shows poor reproduction and hampered regeneration. (5) With no individual in seedling and sapling stages but relatively many individuals in tree/shrub stage (e.g., Ekebergia capensis and Schefflera abyssinica): species exhibiting this pattern were not regenerating. Even though Schefflera abyssinica exhibits this pattern of regeneration, it International Journal of Ecology is difficult to conclude the species was not regenerating. e main reason is that it grows as an epiphyte mainly on other tree species and finally overtakes it to become an independent tree; as a result, the seedling and sapling stages are not visible on the ground [40]. From the three classes established for priority setting based on their regeneration status for the sake of conservation activities, those species under class 1 and class 2 are recommended to be given the highest priority.
us, all stakeholders at both national and regional level should participate in the conservation endeavor of these species which can encompass both in situ and ex situ conservation.

Conclusion
Description of floristic diversity of woody species in Agama Forest revealed the presence of high species diversity. Of the species recorded in this forest, two tree species, Vepris dainellii and Millettia ferruginea, are endemic to Ethiopia. Structural analysis and assessment of regeneration status of woody species in this forest showed that the overall ecological condition of the forest was healthy. However, structural analysis and assessment of regeneration of some species revealed that there are species which exhibit abnormal population structure and abnormal pattern of regeneration which in turn necessitates conservation and management of these species. e in situ conservation strategy, which has been implemented by FARM Africa, an NGO that has been engaged in conservation endeavor of the study forest, in the form of participatory forest management (PFM), should be strengthened via collaboration of all potential stakeholders to reverse the unhealthy population structure and regeneration status of woody species.

Data Availability
Part of the data used in this research are included and attached as Additional Files 1, 2, and 3. us, the data used for this manuscript are available.

Conflicts of Interest
e authors declare that they have no conflicts of interest.