Vegetation structure and floristic composition of Gergeda Anfillo Forest , West Ethiopia

Ethiopia is a tropical country with tropical, subtropical and temperate types of vegetations. However, many vegetation types are not described and characterized. The aim of this study was to assess the floristic composition and structure of the protected Gergeda Anfillo Forest vegetation. The forest was located in Kellem Wollega Zone, southwestern Ethiopia. Stratified sampling method was used for vegetation data collection. Samples were taken from 40 quadrats of 20 m x 20 m (for woody species) that were laid along transect laid about 200 m apart. 120 subplots (1 m x 1 m) were used for herbs in the main plots. A total of 134 (11 endemic) species, 116 genera, and 61 families were recorded. The five most dominant tree species with highest importance value index were Schefflera abyssinica, Ekebergia capensis, Albizia gummifera, Croton macrostachyus, and Olea welwitschii. The most abundant families were Fabaceae (14 species), Poaceae (11 species), Asteraceae (10 species), and Euphorbiaceae (7 species). Moreover, the forest housed 10 of the 24 national priority tree species and four plant communities were identified by cluster analysis. Structural analysis revealed that the forest is dominated by small sized trees and shrubs. This indicates that the forest is facing selective cutting and/or it is in secondary stage regeneration. Given the non-seasonal climate, less differentiated communities, and high plant diversity, it is possible to conclude that the forest can be categorized as tropical rainforests of western Ethiopia. Thus, protection of the forest is highly recommended.


INTRODUCTION
The floristic diversity of Ethiopian is appreciated for having more than 800 endemic species, center of origin and/or diversification of globally important crops (IBC, 2012).
However, the country is frequently alighted for the diminishing vegetation cover.Because of increasing demand for farmlands, fire wood and livestock feed there was a widespread deforestation (Chaffey, 1980;Million Bekele and Leykun Berhanu, 2001;Feyera Senbeta and Demel Teketay, 2003;Teshome Soromessa et al., 2004).According to Ensermu Kelbessa et al. (1992), the limited income generation opportunities had forced the Ethiopian farmers to cultivate and graze marginal lands, which catalyzed the environmental degradation vicious circle.
Loss of forest resources would have negative impacts on ecosystem services, biodiversity conservation (Zerihun Woldu, 2008) as well as socio-economic wellbeing's.Ecosystem disruptions have already threatened a number of plant species (Ensermu Kelbessa et al., 1992).Of course, loss of forest cover and biodiversity due to humaninduced activities is an alarming concern in many *Corresponding author: alinabiot@yahoo.comparts of the world (Petman et al., 2001;Feyera Senbeta andDemel Teketay, 2003, Vivero et al., 2006).To alleviate the global and regional losses, due attention has to be given to the establishment of natural reserves (WCMS, 1992;Fashing et al., 2004).
There were many studies of tree community structure and composition which have been conducted throughout the tropics to document and explain the patterns of tree diversity found in the earth's tropical forests (Pitman et al., 2001, Reyers, 2004).One outcome of these studies has been the realization that tropical forest tree community structure and composition varies widely not only between forests on different continents (Phillips et al., 1994), but also between forests on the same continent (Ter Steege et al., 2000) and even between different sites within the same forest (Proctor et al., 1983).

Botanical assessments of floristic composition and
structural studies are essential in understanding the extent of plant diversity and health of forest ecosystem (WCMC, 1992).The availability of up-to-date data on forest resources is an essential requirement for forest management planning and sustainable resources utilization (FAO, 2007).Even though, there are some general vegetation surveys during the NFPAs selection, perhaps because of its relative remoteness of its location, Gergeda Anfillo Forest was not studied before.Though, the forest has been under protection, its plant diversity and vegetation structure has to be studied for proper protection and monitoring.
Thus, this study was designed to make investigations on floristic diversity and related forest attributes of the Gergeda Anfillo Forest.
The objectives were to document the floristic composition; identify plant communities; and to determine vegetation structure of the forest.

Description of the study area
The study area is located in Duli kebele, Anfillo District, Kellem Wollega zone, Oromia Regional State, Ethiopia (Figure 1).2) resembles the tropical rainforests rather than the other parts of Ethiopia.However, the temperature is relatively lower than rainforests.This could be explained by the higher altitude of the study area.
Climadiagram of the study area was drawn using ten years data of the Anfilo Metrological Station (Figure 2).

Vegetation data collection
Reconnaissance survey was conducted from November 8-12/2012 in order to obtain an impression of the site condition and to determine the sampling methods to be used for vegetation data collection.
Stratified sampling method was used as described by Kent and Coker (1992).A total of 40 (20  represent homogeneous vegetation units.The total plots were decided as no new species appear by adding plots (Braun-Blanquet, 1932;Blackman, 1935).
Environmental variables such as altitude and geographic coordinates (latitude and longitude) were taken from the center of each main plot using Geographical Positioning System (GPS).Identification and nomenclature were based on Flora of Ethiopia and Eritrea (Edwards et al., 1995(Edwards et al., , 1997(Edwards et al., , 2000;;Hedberg et al., 1989Hedberg et al., , 2003)).The endemic species and level of threats was based on Ensermu Kelbessa et al. (1992) and Vivero et al (2005Vivero et al ( , 2006)).Diameter at Breast Height (DBH) of all woody plants having diameter greater than two centimeters measured using a meter tape.Height was estimated by Santo-Clinometers and using calibrated stick.

Data analysis
Vegetation classification was done by using SPSSversion 20 software.Every species was given two figures, the first expressing its abundancedominance, and the second its sociability.These lists of quadrats were then tabulated in tables of 'presence' or occurrence, from which the species are classified together according to their affinities and fidelity into communities of different types.
Community names were given after characteristic species that were easily observed in the forest.
Shannon-Wiener (Magurran, 1988) index of species diversity was used to evaluate diversity.
Shannon's Evenness (E) calculated from the ratio of observed diversity to maximum Shannon-Wiener diversity (Kent and Cooker, 1992).The similarity of vegetation types with regard to species composition was assessed using Sorensen's coefficients as described by Grieg-Smith (Kent and Coker, 1992).Density, frequency, height, diameter at breast height (DBH), species importance value (SIV) and basal area were calculated as Kent and Coker (1992) and Muller-Dombois and Ellenberg (1974).
The total basal area was calculated from the sum of the total diameter of immerging stems.In trees, shrubs and saplings, the basal area was measured at breast height (1.5m) and by using the formula Species Important Value (SIV) combines data from three parameters, which include Relative Frequency, Relative Density and Relative Basal area (Kent and Coker, 1992).Importance value index is the most realistic aspect in vegetation study and used to compare the ecological significance of species (Lamprecht, 1989).
The vertical structure of the woody species occurring in the Gergeda Anfillo Forest was analyzed using the International Union Forestry Research Organization (IUFRO) classification scheme (Lamprecht, 1989).The scheme classifies the storey into upper, where the tree height is greater than 2/3 of the top height; middle, where the tree height is in between 1/3 and 2/3 of the top height and the lower storey where the tree height is less than 1/3 of the top height.afromontanum.

Vegetation community classification
The vegetation classification statistics using SPSS-20 revealed four clusters and two outliers (Figure 3).The clusters identified the four plant communities.In all observed plant communities, species with higher relative abundance and relative frequency are those that were easily observable and repeating themselves in associations.Thus the identified groups are more or less coinciding with the natural associations that anyone can observe.The four communities were named after the characteristic tree and/or shrub species and described below.

Species diversity of communities
As it can be seen from Table 2, the study area not only shows species richness but also shows very good Shannon -Wiener Diversity and evenness.
among communities.The highest similarity (least dissimilarity) was observed between communities 1 and 4 (35.2%)followed by community 2 and 4(33.9%) and community 2 and 3(32.7%)due to the communities having close altitudinal similarity and adaptation.Though the similarity coefficients are small (< 50%), they indicate the existence of ubiquitous species with a wide range of tolerance.
The least similarity (highest dissimilarity) was observed between community 3 and 4 (26.4%),followed by community 1 and 2 and community 1

Similarity between plant communities
In order to determine the similarities among plant communities of the study area, similarity ratios were computed following Sorensen`s similarity coefficient (Table 3).Based on this, similarity in species composition slightly varied and 3. i.e., 30.2% and 31.5%,respectively (Table 3); this may be due to conservational variation and variation in disturbance due to anthropogenic activities, i.e., one area which is better protected varies from the one which is highly exposed to deforestation resulting in communities variation.

Vegetation structure a) Vertical structure
The tallest tree observed in Gergeda Anfillo  It is important to note that the highest proportion of species was concentrated in the lower storey (61.96%) followed by the middle (26.07%) and upper storey (11.85%) of the vertical structure of the Gergeda Anfillo Forest (Table 4).Similar result was also observed by Ensermu Kelbessa and Teshome Soromessa (2008) in which few species attained the upper story in Bonga Forest.

b) Vegetation height and DBH classes
The height and Diameter at Berst Hight (DBH) graphs were more or less inverted J-shape

c) Basal area
The total basal area of the forest is 40.94 m 2 ha - 1 .There is a considerable decrease in number of individuals with increasing DBH size and basal area.Individuals that attained higher DBH classes are fewer in number while they contributed over 77.69% to the total basal area.The density distribution of tree species does not follow the patterns of basal area with a polynomial curve fitting R 2 = 0.89 (Figure 5).For example, species such as Schefflera abyssinica and Pouteria adolfifredericii having the highest basal area do not necessarily have the highest density.More or less similar results were observed for natural conditions in Wof-Washa (Tamirat Bekele, 1994) and in Masha-Anderacha (Kumilachew Yeshitila and Taye Bekele, 2003).
Height class could reflect the different growth phases or ages of tree species, the stratification and stage of forest succession.Height is a good indicator of the role of species that determines the vertical structure of the stand (Pascal and Pelissier, 1996).Thus, tree that were found in sampling plots were classified into eight height classes.5).The relative importance of tree species in a forest can better be depicted given species in comparison to other species in the area.It also is used for setting priority /ranking species management and conservation practices.
Furthermore, it helps to identify the sociological status (structure) of species in a certain plant community (Kent and Coker, 1992).

CONCLUSION
The vegetation of Gergeda Anfillo Forest showed

Figure 1 :
Figure 1: Location map of the study area in Amfilo District of the Oromia Regional State

For
drawing the climadiagram of the study area ten years (2003 to 2012) data were collected from the office of National Metrological Agency, Addis Ababa.A complete list of plants was recorded from species falling in quadrates, herbarium specimens collected and percent cover of species visually estimated then converted into cover abundance values by the modified 1-9 Braun-Blanquet scale (van der Maarel, 1979).Vernacular names of species were recorded during field work.

Figure 3 .
Figure 3. SPSS classification Dendrogram showing community types and two outliers (numbers along X-axis are quadrats)

Forest
was Pouteria adolfi-fredericii with 40 m height.Trees in the lower, middle and upper storey were with height range < 13.3 m, 13.3-26.6m and >26.6 m, respectively.The emergent tree species that occupied the upper storey in Gergeda Anfillo include Pouteria adolfi-friederici, Albizia gummifera, Croton macrostachyus, Prunus africana, Olea welwitschii, Ficus sur, Schefflera abyssinica, and Albizia schimperiana.In addition, the upper storey had low ratio of individuals to species ( Figure 4.) with rogation power curve fitting of R 2 > 0.8.When taken together, like the frequency distribution of the DBH classes, the height classes almost attained a regular (normal) distribution pattern except for the last height class that was represented by individuals forming the upper canopy of the forest.More number of individuals per hectare found in lower height and DBH classes, which contributed to larger proportion (42.53%) for height class one.This could suggest that the Gergeda Anfillo Forest is dominated by lower heighted individuals.Such patterns commonly referred to as reverse J-shape distribution showing stable population structures, but there would be variation with respect to individual species when it was analyzed separately.

Figure 4 .
Figure 4. Density (%)/ha of tree/shrub in different height classes (Above) and DBH classes (Left) the glimpse of rich plant diversity of southwestern Ethiopia.The study area had high (>100) species composition.It had also good richness of genera and families.Of the recorded species, 11 are found to be endemic.The forest had four plant communities that were less differentiated along altitudinal gradients, indicating the undifferentiated nature of the forest.The mid-altitude community (community II) exhibited the highest richness while community IV which was found at high altitude showed the highest diversity.Given the non-seasonal climadiagram, high plant diversity, structural complexity and less differentiated communities, it is possible to conclude that the Gergeda Anfillo Forest showed the glimpse of tropical rainforest of western Ethiopia.Moreover, the forest can be considered as an important site for national biodiversity conservation.As the forest is a vital national resource, we recommend the conservation of forest biodiversity.Moreover, further investigations, particularly on ethnobotany, soils, and detailed ecology of the forest, are required.

Table 4 .
Density, Species number, and individuals to species ratios by story

Table 5 .
Top ten plant species in Species Importance Value (SIV) and their rank Key: RD = Relative Dominance; RF; Relative Frequency; RDe =Relative Density & SIV= Species Importance Value