The study was conducted in Sheleko Medhanialem forest, which is found in Fogera woreda, South Gondar Zone, Amhara Region. Fogera woreda is one of the 169 districts of the Amhara Regional State, found in South Gondar Zone. It is situated at 11˚46' to 11˚59' latitude North and 37˚33' to 37˚52' longitude East. Woreta is the capital of the woreda and is located 625 km from Addis Ababa and 55 km from the regional capital, Bahir Dar ( Figure 1 ). The woreda is bordered by Libo Kemkem Woreda in the North, Dera Woreda in the South, Lake Tana in the West and Farta Woreda.

The area has a uni-modal rainfall distribution pattern. The climate diagram of the study area was drawn based on 10 years of data. The mean annual rainfall in the study area was about 1401 mm, and its peak period was from July to October. There was a decreasing rainfall trend from November to December and none or very little from January to May. The mean minimum and maximum temperatures of the area were 21˚C, 13.5˚C and 29.8˚C respectively (Amhara Metrological Service Center from 2010 to 2019) ( Figure 2 ).

Preliminary survey was done on the first week of November 2019 across the forest to get an overview of the forest to select sampling points and determine the sample size. During the preliminary survey, contact was made with relevant and concerned authoritative bodies and with the people living around the forest. The survey was useful in providing a general understanding of the existing conditions of the study forest.

Sampling design

Systematic random sampling design was applied in this study. Sheleko Medhanialem church forest has an area of 250 hectares. Among this area of the forest 20, 000 m² area was sampled. Five parallel line transects were systematically laid at every 200 m distance between consecutive transects in the North to South direction. The first line transect started at 50 meters distance away from the edge of the forest to avoid “edge effect”. A total of 50 sample plots having an equal size of 20 m × 20 m (400 m²) were laid along transect lines, to collect data on matured alive woody species at a horizontal distance of 100 m from each other. Within the main plot 5 m × 5 m subplots were laid four at the corners and one in the middle for seedling and sapling data collection.

Woody plant species data collection and identification

The actual data collection was conducted from December 2019 to January 2020. All woody plant species encountered in each plot were recorded by their local (vernacular) names. Local names of the woody plant species were provided by informants who were born, brought up, and lived around the study forest area throughout their lives. Woody plant species occurring outside the plots but inside the forest within 10m distance from plots were also recorded only as ‘present’, but not used in the subsequent data analysis [27] . To avoid confusion, colored chalk and tally sheet were used to mark and record woody plants respectively. Taxonomic identification was made by referring to the published volume of the flora of Ethiopia and Eritrea [28] as well as personal expertise. For the purpose of regeneration assessment, the seedlings and saplings from the five subplots were counted for each species. In this study, a seedling was considered a woody individual with a DBH < 2.5 cm and a height of 1 - 2 m.

Environmental data collection

The latitude, longitude and altitude were taken from the center of each main plot (Appendix 5) and measured using global position system (GPS) (Garmin 64 GPS). The center of the main plot was found by stretching two ropes from horizontally and vertically stretched ropes and assisted by GPS. Aspect was recorded as N = 0; E = 2; S = 4; W = 2.5; NE = 1; SE = 3; SW = 3.3 and NW = 1.3 using compass. The type of disturbance was also visually evaluated and recorded for each plot. Disturbances were recorded using the scales 1 - 5 based on the Ethiopian Biodiversity Institution (EBI) field manual. Thus 0 = Negligible, 1 = Very light, 2 = Light, 3 = Moderate, 4 = Intensive and 5 = Very intensive.

Vegetation Data analyses

Structural and diversity data analyses

The structure and composition of forests was determined following [29] . The importance value index (IVI) was calculated as R (ni/N) ln (ni/N), where ni is the importance value index of a single species and N is the sum of the IVI values of all species. The IVI of each species was calculated by adding the relative values of their frequency, density, and basal area. Canonical correspondence analysis (CCA) was done to analyze the relationship between the environmental variable and vegetation data by fitting the data into the ordination scatter plot. Data analysis of the Sorensen’s similarity coefficient in species composition of Sheleko Medihanialem forest was made with six other Afromontane forests found in Ethiopia.

Regeneration data analyses

The regeneration status of woody plant species was determined based on the population sizes of seedlings, saplings, and adults. Regeneration status of the forest was analyzed by comparing saplings and seedlings with the matured trees according to [30] , the status was good regeneration, if seedlings > saplings > adults; fair regeneration, if seedlings > or ≤saplings ≤ adults; Poor regeneration, if the species survives only in sapling stage, but no seedlings (saplings may be or = adults); and if a species is present only in an adult form it is considered as not regenerating.

A total of 65 woody plant species in 55 genera and 34 plant families were identified from Sheleko Medhanialem forest (Appendices 1 & 2). Four species were observed outside the plot but within the forest and used to make the complete list of woody plant species in Sheleko Medhanialem forest. These were Eucalyptus camaldulensis Dehnh, Ficus ingens Miq., Olea europaea subsp. Cuspidate (Wall. ex G. Don) Cif and Vernonia amygdalina Del.

The species richness of the forest was higher than some other forests of Ethiopia, e.g. Ambo state forest with 58 species located in South Gondar Zone [31] , and Wogello natural forest with 20 species in North Gonder [32] . On the other hand, the species richness of Sheleko Medhanialem forest was less than that of several other dry Afromontane forests like Woynwuha natural forest with 69 species in East Gojjam [33] , Sesa Mariam forest with 113 species in Northwest Ethiopia [34] . The variations in the species richness of different forests might be due to several reasons like climate, soil PH and variations in the intensities of human caused disturbances. Forests with a high degree of human interference and disturbance due to livestock grazing for prolonged periods show relatively lower species richness than others [35] .

The major families recorded in this study were Fabaceae represented by 14 species (21.54%) followed by Moraceae with 6 species (9.23%), and Euphorbiaceae with 4 species (6.15%), Oleaceae with 3 species (4.62%), Asteraceae, Acanthaceae, Celastraceae, Lamiaceae, Apocynaceae, Malvaceae, Combretacae, Myrtaceae each with 2 species (24.62% each). Each of the remaining 22 families was represented by one species (33.85%) (Appendix 2). The most diverse genus was Ficus with 6 species (9.23% of the total species number recorded) followed by Vernonia, Maytenus, Senna, Senegalia and Vachellia containing 2 species (3.08% each contributing to the species richness recorded in this study) and the rest genera contain 49 species (87.69%). Among the 65 recorded woody plant species 32 (49.23%) were shrubs, 29 (44.62%) were trees and 4 (6.15%) were lianas.

Different studies [36] [37] from his study of Gedo Dry Evergreen Montane Forest, West Shewa Zone of Oromia National Regional State, Central Ethiopia, reported the dominance of Fabaceae, Poaceae and Asteraceae in Afromontane vegetation type, but only family Fabaceae fulfills this report in the study area. The dominance of the families of Fabaceae and Moraceae were agreed with the result of Ambo state forest, 17 and 8 species in each respectively [32] . This dominance could be related to its ecological adaptations to its habitat conditions. The dominance of Fabaceae was also reported in similar vegetation studies [38] . The dominance may indicate the suitability of the environmental conditions for Fabaceae.

The overall density of mature woody species with DBH > 2.5 cm in Sheleko Medhanialem forest was 2202.5 stems∙ha⁻¹. This was classified into seven density classes ( Figure 4 ) 1) < 5, 2) 5.01 - 20, 3) 20.01 - 35, 4) 35.01 - 50, 5) 50.01 - 65, 6) 65.01 - 80, 7) > 80 stems ha⁻¹. Only nine plant species contributed 79.25% of the total density from the density class 7 which was due to the dominance of species Carisa spinarum, C. aurea, M. obscura, G. ferruginea, C. tomentosa, P. schimperi, R. glutinosa, P. stellatum and C. macrostachyus. From density class 1 S. kunthianum, F. sycomorus, Ficus palmata, Dichrostachys cinerea, Acokanthera schimperi, Entada abyssinica, F. sur, P. thonningii and Bridelia micrantha contributed only 0.23% of the total density, and those which are represented with less number of species and thus need conservation attention (Appendix 3).

The distribution of matured woody plant species of the study area was classified into seven DBH classes. Class A = 2.5 - 5 cm, B = 5.01 - 10, C = 10.01 - 15 cm, D = 15.01 - 20 cm, E = 20.01 - 30 cm, F = 30.01 - 40 cm, G ≥ 40 cm. The first class had the highest distribution of species density per hectare (1326.5 ha⁻¹, 60.23%), the second takes (624 ha⁻¹, 28.33%) and the 3^rd, 4^th, 5^th and 6^th takes (196.5 ha⁻¹, 8.92%), (25 ha⁻¹, 1.14%), (21 ha⁻¹, 0.95%), (2.5 ha⁻¹, 0.11%), finally the last class takes (7 ha⁻¹, 0.32%) ( Figure 3 ).

In the present study, the density of trees individuals in Sheleko Medhanialem forest with DBH greater than 2.5, 10 and 20 cm ( Figure 4 ). The density of trees with DBH greater than 2.5 was 1950.5 ha ⁻¹. The number of stems with DBH > 10 cm was found to be 221.5 ha⁻¹ and those with DBH > 20 cm was 30.5 ha⁻¹. The ratio of DBH >10 cm to DBH > 20 cm is very high (7.26).

The density of seedlings, saplings and mature plants of the Sheleko Medanialem forest was 2419.2 ha⁻¹, 1737.6 ha⁻¹ and 2202.5 ha⁻¹ respectively. The total density of woody plant species in 36 sample plots of Chilimo-Gaji Forest was 3328.47 individuals∙ha⁻¹ [38] . In the same study, the total density of seedlings, saplings and trees/shrubs were 1743.7, 827.1 and 757.6 ha⁻¹, respectively (Appendix 5). The density value of seedlings and saplings is considered as an indicator of regeneration potential of the species [39] in which the presence of good regeneration indicates the suitability of the environment to the species, which is in turn affected by climatic factors and biotic interference influence.

Poor regeneration, if the species survives only in sapling stage, but no seedlings (saplings may be <, > or = adults) and if a species is present only in an adult form it is considered as not regenerating [31] . The documented woody species have shown stable population structures composed of the highest density of individuals at the lowest DBH class followed by gradually declining densities of individuals with increasing DBH classes [40] . The presence of seedling recruitment is a building block in the population dynamics of many species of trees [41] . So that it is one of the important points in determining local abundance and diversity of mature trees.

The result showed that there are more seedlings than saplings and mature individuals implying the survival of seedlings to reach the sapling stage and the forest is now in a fair regeneration. The presence of a sufficient number of seedlings, saplings, and trees in a forest indicates successful regeneration [42] . The density of species regeneration is expected to vary spatially because of forest structure and physiographic conditions [11] . However, there are also species with no seedlings and saplings that may be caused by the physical condition of their microhabitat and human impacts, and urgent measurements need to be taken as they are in poor regeneration. This means there is no regeneration potential for the particular plant species or the study forest in general. The causes of failure to regenerate include lack of viable seed production, insect and animal predation, unfavorable microclimatic conditions, overgrazing, habitat changes, and biological invasions. The understanding of processes affecting the patterns of regeneration of forest-forming species is of crucial importance to ecologists and forest managers [43] .

The ratio of DBH > 10 cm to DBH > 20 cm is very high (7.26). Sheleko Medhanialem forest has more trees in lower DBH classes than in the higher classes when compared to Gedo Forest [37] , Chilmo and Menagesha forests [28] . The reasons behind these may be geographical location, the nature of the forest, altitude variation, age of the forest, degree of conservation and exposure to disturbance.

DBH distribution patterns indicate the general trend of population dynamics and recruitment processes for a given species. In this study, DBH class distribution of individuals in different size classes, except in the 5^th class showed an inverted J shape distribution. This means that large numbers of individuals are distributed in the lower classes which later decreases in the following upper classes. Large percentage (60.23%) and (28.33%) of individuals were distributed in the 1^st and 2^nd DBH classes respectively whereas only 11.44% of individuals were distributed in the remaining upper classes. Analysis of some selected species in the study site revealed three general patterns. The first type of pattern was an inverted J-shape. It shows a high number of species in the lower DBH classes and a reduction in the highest DBH classes. This pattern was exhibited by the species Pittosporum viridiflorum Sims.

The second type of population pattern was bell-shaped and is characterized by the species Croton macrostachyus Del.; It shows a fairly high number of individuals of the species in the middle DBH classes but lower numbers of individuals of the species in the lower and higher DBH classes. This species has poor recruitment potential which might be due to intense competition between the other species found in its surroundings and also were use of this tree for making fencing, fuel wood, charcoal, and agricultural tools. The third type of population pattern is represented by Acacia persiciflora Pax.; In such a pattern, the number of individuals increases with increasing DBH up to some point and then decreases with increasing DBH. The pattern continues to decrease to some extent and increase density as DBH increases. This population structure pattern showed an irregular or zigzag type of distribution and is not healthy because of selective removal of the species for charcoal, fuel wood and making agricultural tools.

Successful management and conservation of natural forests requires reliable data on aspects such as the regeneration trends [44] . Forest stands, where there are numerous young individuals and lesser mature ones, are recognized as having an inverse J-curve diameter distribution [45] .

The total dominance of woody species in the forest was 49.78 m²∙ha⁻¹. From recorded woody plant species, the most dominant species in Sheleko Medhanialem forest was A. lahai which contributed 16.13% followed by S. guineense which contributed 10.35% and C. aurea which contributed 7.87%. But the least dominant species were F. sur, Combretum collinum, P. thonningii, Justica schimperiana, E. abyssinica, T. brownii, S. kunthianum, H. mystacinus, F. palmata, A. schimperi, F. sycomorus and D. cinerea which contributed 0.1% of the total dominance. According to [24] the high dominance and/or abundance of a few species in a forest could be attributed to a number of factors, such as the overharvesting of the desired species, disturbance factors, succession stage of the forest and/or survival strategies of the species.

According to the comparisons of this study forest, Sheleko Medhanialem forest, with similar studies (4 dry Afromontane forests in Ethiopia), on the basis of tree densities with DBH > 10 and 20 cm (a), DBH > 20 cm (b) and the ratio (a/b), the ratio a/b indicate that Sheleko Medhanialem forest has more tree in lower DBH classes than in the higher classes when compared to Gedo, The relatively larger value of the lower DBH class of the Sheleko Medhanialem forest than the middle and high DBH classes may indicates that this forest has a very large number of young woody plants (seedlings and saplings), and may show high regeneration potential of the study forest. The most probable reasons for these patterns could be different types. This may include past history of anthropologic disturbances [41] , tree logging efficient species recruitments, forest abandonment (recent decline in the extent of anthropogenic disturbances), within-gap regeneration stages of various ages after tree fall the nature of the forest, altitude variation, age of the forest, degree of conservation and exposure to disturbance. The ratio of density of individuals with DBH > 10 cm to density of individuals with DBH > 20 cm is taken as a measure of the size class distribution Grubb et al. [46] , cited in [47] . Moreover, the inverted J-shape distribution pattern is additional evidence or indicator of showing a higher number of individual’s species in the lower DBH classes and reduction at the highest DBH classes. Such a pattern may indicate a normal or healthy structural pattern with good reproduction and recruitment capacity of a given species [24] . Diameter class distribution of selected tree species demonstrated various patterns of population structure, implying different population dynamics among species [48] . The slight reduction in diameter classes of some species may be due to selective removal of small diameter class individuals either by local dwellers for some purpose (e.g., for fencing and fuel wood), or by livestock (trampling or browsing), or maybe other biotic factors, indicating selective harvesting of individuals in the particular size classes.

The nine most dominant woody plant species in Sheleko Medhanialem forest in Importance Value Index are Acacia lahai (the highest IVI value), Calpurnia aurea, Carissa spinarum, Grewia. ferruginea, Maytenus obscura, Premna schimperi, Croton macrostachyus, Caparis tomentosa, and Rhus glutinosa ( Table 1 ).

The Woody species in the forest were grouped into five IVI classes based on their IVI values (IVI class 5 = <1, 4 = 1 - 5, 3 = 5.1 - 10, 2 = 10.1 - 15, 1 = > 15) for conservation priority (49). The distribution of species in IVI class is shown in Table 2 . Although the number of species in IVI class 5 is greater than the other IVI classes, their contribution to the total IVI was 2.93%. Priority class 5 (Appendix 4) should get the uppermost conservation priority since these species are at risk of local extinction. Those species with lower IVI values need high conservation efforts while those with higher IVI values need monitoring management.

Importance Value Index (IVI) is an important parameter that reflects the ecological significance of species in a given ecosystem [49] . IVI combines data from three parameters, i.e., RF, RD and RDO [49] . The highest IVI shows the extent of dominance of a given species in relation to the other species in the structure of the forest stand. Those species with lower IVI values are much larger in number than the most dominant higher IVI values but very few in number, the latter need high conservation efforts while those with higher IVI values need monitoring mana- gement. Priority class 5 (Appendix 4) should get the uppermost conservation priority since these species are at risk of local extinction.

Species with the highest importance value index are the most dominant of the particular vegetation. Similarly, IVI is used to set priority species management and conservation practices and aids in identifying their sociological status in a certain plant species as dominant or rare [49] . Based on their higher IVI value, the nine dominant and ecologically most significant shrubs and trees species in Sheleko Medhanialem Forest were C. spinarum (29.9), C. aurea (25.63), A. lahai (23.99), G. ferruginea (21.31), M. obscura (20.78), P. schimperi (19.0), R. glutinosa (17.02), C. tomentosa (15.47) and C. macrostachyus (14.67). These species covered 62.59 % of the total importance value index. The possible reason for species that had large importance was their high dominance or density. In contrast to this, species like F. sycomorus, F. palmata, A. schimperi and S. kunthianum were species with low IVI values that needed urgent feedback to regenerate (Appendix 4). The possible reason for this could be either the selective cutting of these species by near community or unfavorable conditions. It indicates the requirement of conservation and management of the forest as a whole.

List of woody plants collected from Sheleko Medhanialem forest; key:* = species found outside the plot, ** = endemic plant)

Plant families with their respective number of genera and species.

Density per hectare of mature tree and shrub species, sapling and seedling, in Sheleko Medhanialem Natural Forest. Key: DMTSs+ = Density of matured trees and shrubs, DS = Density of Saplings, DSE = Density of Seedlings.

Importance Value Index.

Locations of vegetation data collection in Sheleko Medhanialem Forest.