Ethiopia is an important regional biological diversity hotspot due to wide ranges of altitude, and geographical features, such as high and rugged mountains, flat-topped plateaus, deep gorges, river valleys and rolling plains (Kelbessa et al., 1992; Woldu, 1999). These contributed to emergence of a variety of habitats suitable for the evolution and survival of various plant and animal species.

The rich biodiversity resources, including forests, are being destroyed at an alarming rate largely due to human related disturbances (Aynekulu, 2011). North western highlands of Ethiopia have only fragments of natural forests scattered and confined to inaccessible and sacred places (Wassie et al., 2005), which suggested that the highlands were once covered by high forests. Therefore, appropriate and immediate measures are required to maintain and restore the remaining natural forests. Accelerated deforestation that arise largely due to the conversion of forests to other agricultural land-use types and the overutilization of forest resources to satisfy the food and energy requirements of the increasing population are major environmental concerns in Ethiopia (Friis et al., 2001; Teketay, 2001). In the absence of suitable interventions, current patterns of deforestation will decimate the remaining forests in the near future (Teketay, 1992; Bekele, 1993, 1994).

Amoro Forest is composed of patches of dry evergreen Afromontane forests found in north western part of Ethiopia. It has been continuously exploited by surrounding communities for agricultural land, firewood collection and charcoal production and construction material. Prunes africana is highly sought by locals for fuel wood and charcoal production (Personal observation). Detailed description of resource exploitation needed analysis of floristic composition and structural complexity could contribute towards the conservation of vegetation resources. It is additionally important to document the remaining vegetation resources for posterity. Nevertheless, such data on woody plant composition, community and structural complexity are lacking for the Amoro forest in Dega Damot District and hence the current study. This is believed to contribute a lot to the effort being made in the development of a sound management plan for effective conservation of the forest resources in the study area.

Study area

Amoro Forest is located in Dega Damot District, Amhara Regional State, North western Ethiopia (Figure 1). The forest is located between 10°50’06.53'' latitude 37°35’51.94'' longitude. The major town nearby is Feresbet and is far from 3 km from the forest. The district is also characterized by good climate for most of the year with annual rainfall between 900 and 1200 ml. Topographically, it consists of 35% mountainous, 30% ups and downs, 20% valleys and 15% plains. The soil of the district is reddish (clay soil), black soil, sand soil, brown soil and white soil types (Dega Damot Woreda Agricultural Office (DWAO), 2017).

Shannon's evenness index (J) was also calculated using:

where, H' = Shannon–Wiener Diversity Index; and H'max = lns where s is the number of species in the sample.

Structural data analysis

Structural characteristics (stem density, basal area, and DBH class distributions) were calculated for each woody plant species. Generally, the following formulas were used to calculate frequency, density and basal area of woody species.

where DBH is diameter at breast height.

Importance value indices (IVI) were computed for all woody species based on their relative density (RD), relative dominance (RDO) and relative frequency (RF). This index is used to determine the overall importance of each species in the forest system (Kent and Coker, 1992).

The Importance Value Index (IVI) for each woody species was computed using the following formula:

Population structure of woody species was analyzed for the entire forest and each woody species based on defined DBH classes. Diameter was classified into seven classes. DBH classes (cm) were 2.5–10, 10.1–20, 20.1- 40, 40.1-60, 60.1-80, 80.1-100, > 100 cm).

Forest composition

A total of 57 woody species representing 38 families were recorded from 36 plots (Appendix Table 1). Of the total plant species identified, 19 species (33.33%) are trees while 31 species (54.38%) are shrubs. The remaining 7 species (4.37%) are lianas.

The results of this study show that the woody species composition of Amoro Forest (57 species) were higher than many Afromontane forests in Ethiopia and other tropical forest. For instance, Tadele et al. (2014) recorded a much lower species (50 species) in Zengena Forest in Ethiopia, Berhanu et al. (2016) recorded 66 woody species in Kuandisha Afromontane Forest in Ethiopia; whereas Neelo et al. (2015) recorded 47 woody species in open and exclosed dry woodland sites around Molapo farming areas of the Okavango Delta in Botswana. On the other hand, Zegeye et al. (2011) recorded a much higher woody species 143 species in Tara Gedam and Abebaye forests in Ethiopia. The reasons for variation in floristic composition at the study sites could be due to excessive anthropogenic disturbances, disparity in conditions for regeneration and exploitation of some species. Besides this, the geographical location of the study area differs from other areas. According to Chen et al. (2004), environmental heterogeneity, regeneration success and competition are also important factors that shape species composition of forests.

Four plant community types were identified from the hierarchical cluster analysis, Calpurnia aurea-Vernonia myriantha, Vernonia myriantha-Euphorbia abyssinica, Laggera tomentosa-Solanecio gigas, Allophyllus abyssinicus-Bersama abyssinica (Figure 2; Table 1). Unfortunately, this study did not address analyses of a range of possible environmental variables except altitude that could shape the distribution of identified plant communities. The cluster results of four groups of communities in the present study suggest the existence of overlapping altitudes (Körner, 2000).

Calpurnia aurea-Vernonia myriantha type: This occurs between altitudinal 2432 and 2615 m a.s.l. This communitytype is characterized by C. aurea and V. myriantha as the dominant shrubs. Other important species in this group include Clausena anisata, Clutia abyssinica, B. abyssinica, C. anisata, Euphorbia abyssinica, Maytenus arbutifolia and Dovyalis abyssinica.

Vernonia myriantha-Euphorbia abyssinica type: This community type was situated at altitudinal of 2455 to 2622 m a.s.l. This community was dominated by V. myriantha and E. abyssinica. Some of the species are C. anisata, C. aurea, A. abyssinicus, D. penninervium and U. hypselodendron.

Allophyllus abyssinicus-Bersama abyssinica type: This was encountered at an altitudinal range of 2463–2728 m a.s.l. This community type was dominated by A. abyssinicus and B. abyssinica. The tree layer consists of tree species such as P. africana, A. dimidiata and E. abyssinica. The shrub layer includes C. aurea, L. giberroa, M. arbutifolia, D. penninervium and Liana species are Rubus steudneri and Embelia schimperi.

Laggera tomentosa-Solanecio gigas type: This community type extends from 2493 to 2567 m a.s.l. In this community type, L. tomentosa and S. gigas are the dominant shrub.

Shannon`s diversity indices showed that use of the name as described above had the highest species diversity (3.62) followed by Communities 2 and 3 (Table 2). On the other hand, Community 1 had the highest number of species (47) and Communities 3 and 4 the lowest (29). Community 1 followed by communities had the highest evenness value. While Communities 2 and 3 had the least evenness value (Table 2).

Vegetation structure

The density of trees and shrubs with DBH greater than 2.5 cm in the Amoro Forest was 2860.5 stems ha^-1 (Appendix Table 2). The forest had relatively high density compared with that of Zengena Forest in North western Ethiopia (Tadele et al., 2014). On the other hand, the density was low compared to some other dry Afromontane Forest in Northwestern Ethiopia such as Tara Gedam and Abebaye Forests (Zegeye et al., 2011), Kuandisha Afromontane Forest (Berhanu et al., 2016) and peninsula of Zegie (Alelign et al., 2007). The density of the woody species varied considerably in the different forests. This could be attributed to variations in topographic gradients and habitat preferences of species forming the forest, and the degree of anthropogenic disturbances (Whittaker et al., 2003). The density of the desired species reduced due to the devastating effects and lack of afforestation.

The species with the highest frequency value was V. myriantha (94.4%) followed by A. abyssinicus (88.9%), P. africana (83.3%), Brucea antidyssentrica and D. penninervium (75% each) and E. abyssinica (72.2%) (Appendix Table 2). High frequency value represents a wider distribution of the species in the forest. The variation in density and frequency among species may be attributed to differences in site conditions, species characteristics for adaptation, degree of exploitation and conditions for regeneration (Shibru and Balcha, 2004).

The total basal area of woody species in Amoro Forest was 18.5 m² ha^-1 and comparable to other Afromontane Forests in Ethiopia like Kuandisha Afromontane Forest (Berhanu et al., 2016) and Zengena Forest (Tadele et al., 2014). However, the basal area of this forest is small compared to other dry Afromontane Forests in Ethiopia like Tara Gedam and Abebaye Forests (Zegeye et al., 2011), Wof-Washa Forest (Bekele, 1993) Adelle and Boditi Forests (Yineger et al., 2008). The normal basal area value for virgin tropical forests in Africa is 23-37 m²/ha (Lamprecht, 1989). Thus, the basal area of Amoro Forest is low compared to tropical forests in Africa including dry Afromontane Forests found in Ethiopia. This may be due to Amoro Forest being dominated by shrubs. Besides this, cutting down trees and other factors influenced species by reducing the number of stems desired, affected species diversity and their size.

The distribution of woody species showed in decrease from lower to higher among DBH classes (Figure 3). The forest showed a reverse “J” distribution. These results were similar for those of Savadogo et al. (2007) in Tiogo Forest. The density was most abundant at DBH classes less than 10 cm (1318 stems ha^-1), 10 -20 cm (440 stems ha^-1), 20-40 cm (205.556 stems ha^-1), 40-60 cm (60.802 stems ha^-1),60-80 cm (28.086 stems ha^-1) and >100 cm (17.901 stems ha^-1). This pattern indicates that the majority of the species had the highest number of individuals in lower DBH which in turn shows that the forest vegetation has good reproduction and recruitment potential.

Schefflera abyssinica had the highest IVI (41.0) followed by P. africana (30.7), A. abyssinicus (25.4), A. dimidiata (23.3), V. myriantha (17.8), S. gigas (12.3), C. aurea (10.7), E. abyssinica (10.6), B. abyssinica (9.4) (Appendix Table 2). IVI value is an important parameter that reveals the ecological significance of species in a given ecosystem (Lamprecht, 1989).

Population structure

The woody tree and shrub species in Amoro Forest were analyzed in four representative size structure (Figure 4a-d). The first pattern was an inverted J-shaped distribution exhibited by species with high number of individuals in the first and second DBH classes and with gradual decrease of larger sized trees (Figure 4a). This pattern was recorded on B. abyssinica, suggesting adequate seedling reproduction and regeneration (Bekele, 1993). Gause II (Figure 4b) lacked individuals at intermediate DBH classes and indicates the sizes missing and individuals present at the high and lower DBH classes. The species Ekebergia capensis shows this type of pattern. This kind of distribution is observed when there is selective logging of medium sized individual. In my observation, the main reasons for the absence of medium DBH classes were selective cutting for construction, charcoal production timber and firewood.

The J shaped was characterized by a higher proportion of larger individuals with over 40 cm DBH (15.12 stem/ha). The trend decreased towards lower DBH classes and absence of smaller individuals below 40 cm DBH (0 stem/ha). This regeneration pattern was observed in S. abyssinica. Such pattern shows poor reproduction (Bekele, 1993) due to the fact that it is possible most trees are not producing seeds due to age. It also indicates the presence of selective cutting of preferred size classes on juvenile individuals by local. The species with this type of pattern have large individuals that are less competent to reproduce and in a weak position of regeneration status selective cutting or grazing effects on juvenile individuals.

Gause I was represented by P. africana (Figure 4d). The first and second DBH classes having low number of individuals, a gradual increase in the number of individuals towards the medium classes, and subsequently a decrease in number towards the higher DBH classes. This pattern indicates a poor reproduction (Bekele, 1993) and recruitment of species which may be associated with human use selectively harvesting branches of stem for construction material and charcoal production.

The woody plant of Amoro Forest are dominated by shrubs (54.38%). Euphorbaceae was the dominant family with six species (10.52%) followed by Asteraceae family with four species (7.01%). Amoro Forest is mainly a dry evergreen Afromontane Forest which is dominated by small sized tree and shrub species. The general size composition of trees in the forest was inverted J, however there were differences among tree species. B. abyssinica species showed typical healthy inverted, whereas S. abyssinica species J with predominance of adult and absence of juveniles. P. africana and Ekebergia capensis showed Gausian form with absence of intermediately sized individuals attributed to selective harvest. The density of woody species decreases with increasing DBH indicating predominance of small sized individuals in the forest. This implies that the forest is in good state of reproduction.

Therefore, to improve the natural diversity and structure of the forest, to minimize the influence of the surrounding communities and utilize the forest resources sustainably for present and future generation, it is necessary to take measures for protection of the forest human interference and other anthropogenic influences.

The authors have not declared any conflict of interests.

The authors express their deepest thanks to the Agriculture and Rural Development Office of the Dega Damot District as well as to the respective chairpersons. We also thank Addis Ababa University for financial support.