Floristic composition and structural analysis of Gelesha forest , Gambella regional State , Southwest Ethiopia

The study was conducted on Gelesha Forest, Godere District and Gambella Regional State, Southwest Ethiopia to determine floristic composition and structural analysis of the forest. Systematic sampling method was used to collect vegetation data from 60 plots (20 m × 20 m) for trees and shrubs, 5 m x5 m for seedlings and saplings and 1 m × 1 m for herbaceous plants. Vegetation classification was performed using PC-ord software package. All trees and shrubs with DBH >2 cm and height >2 m were measured for diameter tape. A total of 157 species of vascular plants belonging to 127 genera and 54 families were recorded from Gelesha Forest. Of all the species recorded 33.12% were trees, 19.11% shrubs, and 13.38% woody climbers. The dominant families were Euphorbiaceae and Moraceae (13 species each): 11 and 6 genera, respectively. The density of trees in the forest decreases with increasing DBH classes. Six population distribution patterns were recognized. The Forest has a density of 557.09 stems/ ha and a total basal area of 98.87 m 2 / ha. The forest is characterized by high density of trees in the lower class than in the higher. Phytogeographically, it was more related to the moist evergreen Afromontane forests of the Southwest Ethiopia than other forests in the country.


INTRODUCTION
A huge proportion of the planet's land surface has been transformed in other land use systems.Although landuse practices vary greatly across the world, their ultimate outcome is generally the same: the acquisition of natural resources for immediate human needs, often at the expense of degrading environmental conditions (Foley et al., 2005).Even though deforestation and forest fragmentation is a global phenomennon, it is severe in the tropica.
Tropical forests are deteriorating both in quality and quantity.Approximately half of the tropical closed-canopy forest has already been removed and the land converted to other uses (Wright, 2005).The rapid conversion of tropical forests for agriculture, timber production and other uses has generated vast, human-dominated landscapes with potentially dire consequences for tropical biodiversity (Gibson et al., 2011).*Corresponding author.E-mail: Kitessah2@yahoo.com.Tel.+25147111012.1 Author(s) agree that this article remain permanently open access under the terms of the Creative Commons Attribution License 4.0International License Ethiopia is located in the horn of Africa between 3°24´and 14°53´ N and 32°42´ and 48°12´ E with a total area of 1,120,000 km 2 (MoA, 2000).Its physical features, range from the depression at Dallol in Afar, which is about 110 m below sea level to the highest peak at Ras Dashen, 4620 m above sea level (EFAP, 1994).Plateau of Ethiopia with an altitude of above 1500 m a.s.l.covers about 40% of the country (Teketay, 2004).Ethiopia has a wide range of habitats due to the different topography of the country.The variation of the physical features coupled with other environmental factors has contributed to the different vegetation types in Ethiopia (Woldu, 1999).The flora of Ethiopia is estimated to be about 6,000 species, of which 10% is considered to be endemic (Hedberg et al., 2009).In Ethiopia, forest cover has been declining steadily.About 35-40% of the Ethiopia's land area was covered with high forests by the turn of 19 th century (Breitenbach, 1963).EFAP (1994) indicated that 16% of the land area of Ethiopia was covered with forests in the early 1950's.This number was reported to decline to 3.1% in 19823.1% in , 2.7% in 19893.1% in and less than 2.3% in 19903.1% in (EFAP, 1994)).Most of the remaining forests of Ethiopia are restricted to the south and southwest parts of Ethiopia, which are less accessible, and less populated (Yeshitela and Bekele, 2003).Even the remnant natural forests in these areas are continuously threatened by human activities (Yeshitela and Bekele, 2003).Different factors are responsible for the declining of forest cover in Ethiopia such as forest clearing for cultivation, overgrazing by domestic animals, exploitation of forests for fuelwood and construction materials all coupled with high rate of human population growth (Senbeta, 2006;Friis, 1992).In addition lack of proper policy framework is also contributing to the decline of forest cover in the country.The estimates of annual rate of deforestation vary from 150,000/ha -200,000/ha (EFAP, 1994).The Southwestern Ethiopia is rich in remnants of natural forests, but they are disappearing at an alarming rate because of encroachment by agricultural activities and the pressure from investors who are converting the forests to coffee and tea plantations (Bekele et al., 2002).
Gelesha Forest is one of the foresat priority areas (FPA) in Ethiopa with total area coverage of about 12,000 hectares (GDARDO, 2011).The Gelesha Forest is under serious threat due to rapid human population growth, the demand for new settlement area and the expansion of investment for coffee and crop cultivation (Woldemariam and Senbeta, 2006).It has been continuously exploited by the surrounding people for agricultural land expansion, timber harvesting (logging), firewood collection and charcoal production, wood cutting for construction and other purposes.
Therefore, the current study was initiated to assess the floristic composition, structural analysis and to evaluate the status of the forest in comparison to other studied natural forests in Ethiopia to draw the attention of policy makers to this biodiversity assemblage to undertake appropriate conservation measures.

Description of the study area
Gelesha Forest is located in Godere District, Gambella Regional State, Southwest Ethiopia (Figure 1).The district is located between 708'-723' latitude and 3452'-3525' longitude.It also has an altitude ranging from 500 to 2400 m above sea level, with the natural forest area in the range of 500-1500 (Friis, 1992).The climate of the area is a hot and humid type (Woldemariam and Senbeta 2006), and Meteorological data obtained from National Meteorology Service Agency (Addis Ababa) indicate that Godere District receives high rainfall between mid-March to October and low rainfall from November to February.The highest annual mean rainfall of the study area within twelve years (1998)(1999)(2000)(2001)(2002)(2003)(2004)(2005)(2006)(2007)(2008)(2009) is 2726.7 mm.The lowest mean monthly temperature is 13.2°C and the highest mean monthly temperature is 33.1°C with an average temperature of 22.2°C.
The soil of the district is reddish brown, well drained and clay in texture (Yeshitila, 2008).

Sampling design
Systematic sampling technique was used for the study.Sampling sites were arranged along transects in different directions according to the topographic nature of the study area.A plot of size 20 m x 20 m (400 m 2 ) was used for trees and shrubs, and a 5 m x 5 m (25 m 2 ) plots for seedlings and saplings and 1 m x 1 m (1 m 2 ) for herbaceous plants.Transect were placed 400 m apart and were laid systematically at every 50 m altitudinal range along transect lines.

Data collection
A complete list of trees, shrubs, woody climbers and herbs were collected from the systematically established plots along each transect.Species occurring within 10 m distance from the plots boundaries were also recorded as present for floristic composition.Plant specimens were collected following standard herbarium techniques.Specimen identification was conducted at Jimma University Herbarium using the Flora of Ethiopia and Eritrea, and the National Herbarium (ETH), Addis Ababa University by comparing with authentically identified specimens.
In each plot, height and diameter at breast height (DBH) were measured for all individual trees and shrubs having DBH greater than 2 cm using a diameter tape.If the tree branched at breast height or below, the diameters were measured separately for the branches and averaged.Trees and shrubs with DBH less than 2 cm were counted.Altitude and geographical coordinates were also measured for each plot using Garmin 72 GPS.

Plant community analysis
Cluster analysis was used to repeatedly cluster the plots into groups, based on the abundance of the species (the number of individuals).Sørensen (Bray-Curtis) was taken as distance measured and Flexible β as group linkage model with a flexible β of -0.25 was used on the vegetation data.For three up to six groups this was executed repeatedly (McCune and Mefford, 2006).After clustering, Indicator Species Analysis (ISA) and a Monte Carlo test were applied to each group to determine the indicator species for  , 1997).For each run, the indicator value for each species and the overall average P-value were calculated.This overall P-value was used to decide the most informative number of clusters and avoid creating additional groups with little improvement to overall significance.The last cluster step adding >0.05 significance to the overall average P-value was selected as the most informative number of groups.Multi response permutation procedure (MRPP) was used to test differences in community composition between groups.The test uses Sørensen (Bray-Curtis) distances as well as a natural group weighing factor ni/Σni (with ni the number of sample plots in each group).

Structural data analysis
All individuals of species recorded in the 60 plots were used in the analysis of vegetation structure.The diameter at breast height (dbh), basal area, tree density, frequency and important value index were used for description of vegetation structure.The structural analysis were performed using Microsoft excel 2007.

Phytogeographical comparison
Phytogeographical comparison of Gelesha Forest with other forests in Ethiopia was conducted based on woody species composition.The similarity index used for this comparison was Sorensen's similarity coefficient (Ss).

Ss = 2a/(2a+b+c)
Where, Ss=Sorensen's similarity coefficient; a= is number of species shared by the two forests; b= is the number of species in forest; c= is the number of species in forest (Kent and Coker, 1992).

Plant community classification
Four plant community types were identified from the cluster analysis and were named after two species that had higher indicator values.The differences among the groups were tested with multiresponse permutation procedures (MRPP) based on Sørensen (Bray-Curtis) similarities.The test statistic T value for the four communities was -19.725570 at (P < 0.001) which show a highly significant difference between the communities.The agreement statistic A was 0.13.The agreement A describes within group homogeneity, and falls between 0  and 1.When all items within groups are identical, A=1 and 0 when the groups are heterogeneous.A plant species with a significant indicator value at P < 0.05 was considered as an indicator species of the group (Table 2).
The description of each of the four communities is given below.

Argomuellera macrophylla -Manilkara butugi community type
This community type was distributed between the altitudinal range of 1010 and 1153 m a.s.l.It was represented by 30 plots and 58 associated species.A. macrophylla and M. butugi were the dominant and main indicator species of the community (Table 2).Rinorea ilicifolia, Milicia excelsa and Margaritaria discoidea were the dominant associated tree species.Asplenium sandersonii and Solanum terminale were associated with the ground layer.

Polyscias fulva -Vernonia amygdalina community type
This community type was situated between the altitudinal ranges of 1060-1082 m a.s.l and contains 4 plots and 19 species (Table 2).

Baphia abyssinica -Antiaris toxicaria Community type
This community type was situated at altitudinal range from 1020-1130 m a.s.l and represented by seven plots and 58 associated species.B. abyssinica and A. toxicaria are dominant tree species in the community (Table 2).The associated dominant tree species in the community includes Trilepisium madagascariense, Pouteria alnifolia, Croton macrostachyus, Albizia grandibracteata, Ficus thonningii, Elaeodenron buchananii, Ficus asperifolia, Morus mesozygia and Ficus mucuso.The herbaceous layer was mainly covered by Olyra latifolia and Oplismenus hirtellus.

Blighia unijugata -Lannaea welwitschii community type
This community type was distributed between the altitudinal range of 1021 and 1154 m a.s.l and contains 19 plots and 71 associated species (Table 2).This community was characterized by three indicator species; B. unijugata, L. welwitschii and Alchornea laxiflora.Other associated tree species include Croton sylvaticus, Erythrococca trichogyne, Mimusops kummel, Dracaena fragrance, Celtis african, Ficus umbellata, Trichilia dregeana and Rothmannia unceliformis.Hippocratea pallens is common climber in the community and Oplismenus hirtellus was the herb which covered the ground layer of the community.

Tree density
For the description of the structure of Gelesha Forest, 45 woody species with individuals having DBH greater than 2.0 cm and height greater than 2.0 m were used.Tree density with DBH greater than 2 cm in the study area was 557.09 individuals/ha and the number of stems with DBH >10 cm was 315.42 and those with DBH >20 cm was 244.58 individuals/ha.
The ratio of density at DBH > 10 cm to DBH > 20 cm is taken as a measure of the size class distribution (Grubb et al., 1963).So, the ratio of individuals with DBH >10 cm (A) to DBH > 20 cm (B) was 1.29 for Gelesha Forest (Table 3).The density of woody species of Gelesha Forest was compared with eight other forests in Ethiopia regarding the ratio of the density of woody species with DBH > 10 to DBH > 20 cm given (Table 3).
Ratio of the density of individuals with DBH greater than 10 cm to individuals with DBH greater than 20 cm in the Gelesha Forest was less than that of Chilimo and Menagesh (Bekele, 1993), Marsha-Anderacha (Yeshitela and Bekele, 2003), Dindin (Shibru and Balcha, 2004), Mana Angetu (Lulekal, 2005), Bibita (Denu, 2007) and Magada (Bekele, 2005), showing that all have higher proportions of small-sized individuals than Gelesha Forest indicating that these forests are in the stage of secondary succession.
On the other hand the ratio in Gelesha Forest was more or less similar with that of Dodola (Hundera et al.  2007) indicating that these forests are in climax stage compared with the other forests.

Diameter at breast height (DBH)
The distribution of trees in different DBH classes is given in Figure 2. The density of trees from lower to higher DBH classes showed a decreasing trend.Percentage distribution of trees in different DBH classes in Gelsesha Forest was compared with other forests in Ethiopia (Table 4).
In Gelesha Forest, number of individuals in the lower DBH class (I) was smaller than that of Wof Washa, Menna Angetu, Menagesha, Chilimo, Belete, Denkoro, Magada and Bibita forests.With regard to DBH class (II), the percentage of Gelesha Forest was smaller than that of Denkoro, Magada, Wof Washa, Menagesha, Chilimo and graeter than that of Bibita and Belete but similar with   4).In the DBH class (V) the Gelesha Forest was almost equal to Magada but greater than the rest.This shows that in Gelesha Forest the DBH class from 2-20 cm is about 54.59% indicating that the Forest is characterized by smaller trees than others or showing an inverted J-shape (Figure 2).

Basal area
The basal area of all tree species in Gelesha Forest was 98.87 m 2 /ha.The normal basal area value for virgin tropical forests in Africa is 23-37 m 2 /ha (Lamprecht, 1989).Thus the basal area value of Gelesha Forest is very high.The highest proportion of basal area in the forest was contributed by Lannea welwitschii (22.56%) followed by Baphia abyssinica (15.20%),Celtis zenkeri (12.42%) and M. butugi (8.87%) (Table 5).The trees with highest densities such as Argomuellera macrophylla, Dracaena fragrance and Hippocratea pallens did not contribute much to the total basal area of the forest.
Gelesha Forest is almost similar to Menna Angetu and Wof -Washa Forest, lower than Dodolla, and higher than all the other forests under comparisons with respect to basal area per hectare (Table 6).There is close relationship between basal area and DBH because basal area is calculated from the DBH value.This may be due to variations in the conservation of the forests, exposure to deforestation and geographical location of the forests.

Importance value index (IVI)
Importance value index combines data from three parameters: RF, RD and RDO (Kent and Coker, 1992).It is crucial to compare the ecological significance of species (Lamprecht, 1989).It was also stated that species with the greatest importance value are the leading dominant in specified vegetation (Shibru and Balcha, 2004) 8).
The leading dominant and ecologically most dominant species might also be the most successful species in regeneration, pathogen resistance, preference by browsing animals (least preferred), attraction of pollinators and attraction of seed predators that facilitate seed dispersal within the existing environmental conditions (Kenea, 2008).

Phytogeographical comparison
The result obtained shows Gelesha Forest is one of the most diversified moist evergreen montane forests in the country containing a minimum of 157 species of plants.Direct comparison of the species diversity with other forests is not feasible due to differences in size of forests, survey methods and objectives of the study (Woldemariam, 2003).However, the overall species richness of the vegetation can give more or less a general impression of their diversity and phytogeographical similarity.Accordingly, Gelesha Forest was compared with 11 forests in the country to know the similarity of species in the forests and indicate to which forest type it is related (Table 9).
The overall similarity indices between Gelesha Forest and the other 11 forests in Ethiopia ranged between 0.063 and 0.41(Table 9).Gelesha Forest shared the highest floristic similarity with that of Masha Anderacha (0.41), Yayu (0.4), Bibita (0.35), Sese (0.3), Boginda (0.29) Gurra Farda (0.26), Chato (0.24) and Belete (0.22).The relatively high floristic affinity of the mentioned forests with Gelesha Forest arises from geographical proximity, similarity in altitudinal range, and climatic zones.Geographically, Masha Anderacha, Bibita, Sese and Boginda Forests are located in the Southwestern part of the country.Thus, proximity of the areas could be one of the reasons for such a high floristic similarity.The altitudinal range of Gelesha Forest (1006-1154 m a.s.l) falls within that of the above forests which also has contributed to the observed similarity.Again Masha Anderacha which is found in the southwest part of Ethiopia has an altitudinal range that includes those of Gelesha.On the other hand, Menagesha Suba Forest is found on higher altitudinal range (2600-2750 m a.s.l) and its observed lowest floristic affinity; this is due to high range of altitude and other environmental factors like rain fall, temperature soil.The dissimilarities between forests may arise from the different sample sizes and methods of the study, altitudinal differences, degree of human impact (anthropogenic) action, over grazing and climatic conditions.

Conclusion
Results of the present study shows that 157 plants species were recorded in Gelesha Forest belonging to 127 genera and 54 families.Euphorbiaceae was the dominant family with 13 species and 11genera followed by Moraceae with 13 species and 6 genera, Fabaceae with 12 species and 9 genera and Asteraceae, with 12 species and 8 genera of all the species recorded trees, 33.12% shrubs, 19.11% climbers, 13.38% and 34.39% herbs.Four plant communities were identified and described with varying degree of species richness, evenness and diversity.Variability in population structure, which implies vegetation dynamics, was recognized from the analysis of tree species DBH.In addition, density class description of the forest indicated the dominance of small sized individuals asserting Gelesha Forest is in a stage of secondary development.
The density of woody species in Gelesha Forest decreases with increasing height classes and the forest is characterized by high density of trees in the lower class than in the higher.Thus, the forest is in good state of recruitment.The Forest exhibited the highest BA that is 98.87 m 2 /ha.Three layers (lower, middle and upper) of tree canopies were identified from the study of vertical stratification of Gelesha Forest.Large proportion of woody species was found in the lower storey 48.48% and middle storey contains 31.42%, while upper storey was contains of which 20.04%.
Phytogeographic descriptions and comparisons of Gelesha Forest with other similar forests showed that it is related more to the moist evergreen montane forests than others as it is characterized by dominant species of moist montane forests.

Recommendations
To promote Gelesha Forest conservation, local communities and other concerned bodies should plant indigenous species on all sides of the forest as it was surrounded by farm lands that may serve as a buffer zone.
The species with low important value index should be given appropriate attention and conserved in-situ through the collaboration of local communities, the District Agriculture and Rural Development Office, other interested individuals (Like NGOs) and other stakeholders.

Table 2 .
Indicator Values (% of perfect Indication, based on combining Relative Abundance and Relative Frequency) of each species for each of the four groups and the Monte Carlo test (P*) of the significance observed for each species (bold values indicate indicator species at P*<0.05).

Table 3 .
Comparison of tree densities with DBH >10 and 20cm from Gelesha natural forest with other forests.

Table 6 .
Comparison of Gelesha natural forest, with other 13 afromontane forests in Ethiopia with respect to basal area per hectare.

Table 7 .
Frequency distribution of selected woody species in Gelesha natural forest (Freq = frequency, %FR = frequency percent and RFR = Relative frequency.Species are arranged in decreasing order of frequency).

Table 8 .
The importance value index (IVI) and priority class for conservation of tree species in Gelesha natural forest.

Table 9 .
Comparison of Gelesha natural forest (Godere District) with other 10 forests in Ethiopia based on their similarities/differences Forest.
Ss = Sorenson`s similarity coefficient.a =is number of species shared by the two forests; b= is the number of species in other forest.c= is the number of species unique to the present study.