The Impact of Fencing on Regeneration, Tree Growth and Carbon Stock in Desa Forest, Tigray, Ethiopia

Dryland forests of Ethiopia are facing a rapid rate of deforestation and degradation. This affects the growth and survival of plant density, diversity, growth and carbon stock. The aim of the study was to investigate the effect of fencing on regeneration, bonsai growth and carbon stock potential of Desa‟a forests. Data were collected on 24 plots of 20m×20m for trees and shrubs, 5m×5m for seedlings and on 40m×40m for bonsai growth on permanent plots established in 2009. Data were analyzed using paired t-test to estimate any change on regeneration and carbon stock within the fenced and unfenced block over time. However, an independent t-test was used to measure any significant difference on regeneration, carbon stock and bonsai growth between fenced and unfenced. Results showed that a total of 27 woody species were found representing 18 families. Olea europaea subspecies cuspidata and Juniperus procera were with the highest Importance Value Index. The overall diameter distribution showed an inverted J-shape indicating active regeneration status. Fencing has a significant effect on the regeneration of seedlings (P=0.007). Significant change in the regeneration of woody plants were found within the fenced plots overtime (P=0.001). Unfenced has a significant effect on basal area over time (P=0.036). Bonsai height and diameter growth was significantly affected by fencing (P<0.001) showing mean height and diameter growth of 14.4cm and 1.34cm per year in the fenced plots respectively. Total tree carbon stock of the fenced and unfenced plots was increased from 15.88 to 19.43 and 12.05 to 12.95 Mg C ha-1 respectively over time. Fencing had a significant effect on tree carbon stock over time (P=0.005). In general, long term fencing was found to be an effective management option to enhance regeneration and carbon stock of the forest. Particularly, fencing is very important management options to improve the regeneration of Olea europaea and bonsai growth which is reported to have low regeneration record on an open environment. However, the regeneration potential of Juniperus procera is weak on both management options indicating the need of further study and conservation priority in the dry Afromontane forest environment.


Background and Justification
Dryland forests are among the most and widespread threatened ecosystems from expanding human populations and climate changes, and conversion mostly to agricultural lands [1,2]. Browsing with animals are one of the main challenges for today and tomorrow in forest management and as this it affects regeneration, abundance and forest species distribution Ambroz et al. [3]. Interference of different animals (i.e. disturbances and continuously over browsing) to the forest area affects the growth and survival of many herbs and tree species, as a result plant cover density and diversity is reduced Jean-Pierre [4]. Disturbance and browsing are main factors that lead to change the vegetation and diversity patterns through time in forests worldwide yet rarely are these processes examined both together simultaneously, even though interactions likely to be greater Whigham [5]. In the dryland forest of the world, various forces may influence vegetation dynamics such as regeneration, composition and structure Pickett et al. [6]. According to the study of EFAP [12]; Mengistu et al. [13,14] from the total area of the Ethiopia, 1.1million square kilometer 66% (75 million hectare) is dryland and from these 2.5million hectares of the land is covered by bush lands and wood lands. This Figure 1 indicates that majority vegetation resource of Ethiopia is found in the drylands. Thus, the development of the country is highly dependent upon the efficient and effective use of these areas (UNDP) [15]. In the dry Afromontane forests, soil moisture, light, temperature, nutrient and intensity of browsers significantly affect the seed and seedling quality of the forest composition. The highland area of Ethiopia was once covered with widespread forest resources considered 40% the total are of the country in 1900 EMA, [16,17]. However, these forests are being subjected to intensive anthropogenic pressure, through activities such as cutting, burning and browsing by livestock Birhane et al. [18,19].
Such process can result in deforestation, degradation, clearances, fragmentation of the forests and biodiversity loss Lemenih et al. [20]. In Ethiopia, the rate of deforestation is estimated to range from 150,000-200,000 ha EFAP [21] and recently studies indicate that the net annual loss of forest resources of the country is estimated 210,000 ha by deforestation and forest degradation MEF [22]. The most important reason behind the rapid deforestation and forest degradation rate in the country is the ever increasing human population growth and is associated with a very high demand for agricultural and grazing lands, firewood, charcoal, timber, and construction among others Yineger et al. [23]. Fire, inappropriate investment activities, and lack of viable land use policy have also been key factors for the rapid decline of forests in the country Friis [24,25]. Thus now a days, the forest area is shrinking and left a small remnant forests only found in spiritual churches, monasteries, and mosques and inaccessible hillsides Birhane et al. [26].
Desa"a dry Afromontane forest is mostly dominated by the species of Olea europaea, subspecies cuspidata and Juniperus procera [24,27]. A study by Aynekulu [28] showed that native tree species natural regeneration potential of the forest is low ( Figure   2). Exclosure was found to be an effective management option to improve the regeneration of Olea europaea sub species cuspidata but it does not improve the regeneration of Juniperus procera Giday [27]. Thus, a closed management system in the open-access and degraded forests may not guarantee a successful regeneration of native woody species Aynekulu [28]. There are lots of trees that their growth is restricted vertically as well as horizontally to the

3/17
side due to the frequently browsing effect of many grazing animals both the wild and domestic animals Giday [27]. This is particular to Olea europaea subspecies cuspidata and which results in a persistent reduced and stunted growth called it Bonsai Giday [29]. Protected forests and exclosures in Ethiopia play an important role in carbon storage on their carbon pools in addition to other ecosystem services Vreugdenhil et al. [30]. By absorbing carbon and storing it in trees, shrubs and soil, vegetation plays an important role in reducing the level of greenhouse gases in the atmosphere Australian Greenhouse Office [31]. The carbon pools in these ecosystems are living biomass of trees, understory herbs, the dead mass of the litters, woody debris, soil organic matter and peat. The carbon stored in the above ground living biomass is the largest pool Vreugdenhil et al. [30]. Therefore, maintaining and enriching these ecosystems is a very crucial step wards to a sustainable development and green economy of a nation.
Therefore, the objective of this study is to investigate the vegetation regeneration, tree growth, composition and above ground biomass and carbon stock potentials of the vegetation in Desa"a forest from the permanent plots (Figures 3 & 4).  Basement are the dominant parent materials that are found on the plateau Asrat [33].
Soil: Desa"a forest has a diverse geological formation that plays a big role in soil variability (Asrat [33]. The soil types of Desa"a forest were classified as freely draining and variably textured Phaeozemes, Leptosols, Regosols, Calcaric combisols and Fluvisols are occurred Aerts et al. [34].

Topography
Altitude: Topographically the area exhibits moderately gentle to steep scarps. It forms a climatic buffer zone between the highlands of Tigray Region and the lowlands of Afar Region by extending in an easterly direction along and down the northeastern escarpment. In general the forest lies in an elevation range between 1500 and 2862m.a.s.l Hishe et al. [35]. Specifically the altitude of the permanent plots (research block) ranges from 2480m.a.s.l to 2540m.a.s.l Giday [29] with average altitude 2506m.a.s.l. All the permanent plots of the study area are homogenous in their environmental characteristics in terms of their slope, altitude, aspect, soil depth, stoniness and rock outcrop Giday [27].

Climatic Conditions
Temperature: As the data from (1980 -2016) collected from National Meteorological Agency Atsbi Wenberta station (130 52.7"N and 130 44.6"E) located adjacent to the study area indicate that the mean minimum and the mean maximum temperature for the area varies in the range of 6 0 C to 12.1 0 C and 17.8 0 C to 22.7 0 C, respectively ( Figure 2).
Rainfall: According to National Metrological Agency rainfall data of Atsbi Wenberta district that is near to the study area from 1980-2016, the rainfall of the area is frequently observed during the months of July and August. The amount of rainfall during these months ranges from the 158mm in the month of July up to the maximum 230mm in the month of August. This amount is very low when we compared to that of the other parts of the Tigray region.
Such an amount of rainfall is also marginal for the growth of trees.
However, there is an argument that the forests collect (absorb) moisture from the clouds or moist which usually builds up along the escarpment EWNHS [36].

Vegetation Cover:
The vegetation cover of the forest is mostly dominated by the species of Olea europaea subspecies cuspidata and Juniperus procera followed by the thorny and shrub species of Carisa edulis, Maytenus obscura Giday [29]. There are also few species occurred like Rus vulgaris, Rus natalensis and Dodonea viscose.

Reasons for Selecting the Study Area
The forest area was selected for a long term study by Mekelle University as with permanent plots established in 2009 for the following reasons., One, because it is one of the remaining patch forests under heavy challenge that found in the Tigray region; two, it possess diverse landscape and biological habitat; three, because of the ecological and economic importance it furnishes to the people near and around the forest areas.

Data Collection
Vegetation Composition and Tree Carbon Stock: The nested plots are to increase measurement efficiency, time and minimize counting errors. All trees and shrubs (diameter at stump height (DSH)≥1cm) encountered within the 20m x 20m were identified and measured for either their diameter at breast height or diameter at stump height using diameter tape and height using graduated stick for shorten trees and clinometers for tall trees. All seedlings and saplings (< 1cm diameter at stump height (DSH)) encountered within the 5m x 5m sub plots were also counted and identified KYOTO Giday [27,37]. The trees, shrubs and seedlings were identified and categorized based on local classification knowledge and crosschecked with scientific names on the use full trees and shrubs for Eritrea Bein et al. [38] and trees and shrubs of Ethiopia Bekelle [39] books.

Bonsai Growth
All Bonsai growths, persistently browsed trees, encountered within the 40m x 40m plots were measured for both their diameter and height. Height of individual species was measured using meter tape and graduated stick based on their height accessibility and diameter tape and caliper for the diameter at breast or stump height

Data Analysis
Regeneration and Composition of the Forest: After having tree inventory and measurement, data were assessed for regenerations and structural differences between paired fenced (Exclosure) and adjacent non-fenced plots. The density and percentage frequency of all woody species were calculated and the trees and shrubs were classified based on their DBH classification in to nine and height classification in to eleven classes and then finally drawn a histogram for both diameter and height classes.

Importance Value Index (IVI)
Importance value index (IVI) is used as a measure of species composition that combines frequency, abundance and dominance importance values [40,41]. The ecological importances of tree species in the permanent plot (IVI) were calculated using relative density, frequency and dominance parameters of each species. IVI = RD + Rd + RF …………… Kent et al. [40,41] Relative Density (Rd): Relative density is the study of numerical strength of a species in relation to the total number of individuals of all the species and calculated as Kent et al. [40,41].
Relative Density (RD)= variable (fenced and unfenced management) to estimate the effect of fencing on regeneration was analyzed using independent t-test.
The present data were only used to see the effect of fencing on bonsai growth. Hence fencing was the independent variable and bonsai growth was the dependent variable. The difference was tested using independent t-test. Hence fencing was the independent variable and bonsai growth was the dependent variable. The difference was tested using independent t-test.

Carbon Stock Status of the Forest Above Ground Carbon Stock
In the permanent plot for woody species an allometric equation developed by Chave et al. [43] was used. The development and use of country, biome, climate and species specific equations improve accuracy, minimizes error propagation and reduces bias arising from the generalizations producing by using a generic equations [43,44]. Chave et al. [43] allometric equation was selected because it is more robust than the other generalized allometric equations as it includes diameter, height and specific wood density rather than diameter at breast height only Huy et al. [45]. The inclusion of species and country specific wood density is an important predictor and it improves significantly biomass estimation Chave et al. [43].
For this study specific wood density of the species found in the study area for both indigenous and exotic species done by Ethiopia national forest reference submission report to UNFCC Anonymous [46] was used.  [46]. Where; AGB: Above Ground Biomass of the tree (kg) DBH: is diameter at breast height in (cm) at 1.3m height and H: is Tree height (m) WD: specific wood density of the tree we can take the specific wood density if not available we use the overall average wood density for the species which is 0.612g cm -3 (Anonymous [46]. Species specific allometric equations developed by different authors were used to estimate the biomass of a trees and shrubs. The species specific allometric equations developed by different authors are listed below in a Table 1.    Abundance, Density, Basal Area and Important Value Index:

Forest Composition, Regeneration Status and Population
In 2016, total density of all seedling woody species was found to be 50666 and 1700 stems per hectare in the fenced and unfenced permanent plots respectively (Table 3). Whereas, the density of all seedlings that found in 2009 were 34900 and 24100 stems per hectare respectively (Table 3). Total density of all tree and shrubs woody species in 2016 was found to be 1708 and 1410 stems per hectare in the fenced and unfenced permanent plots respectively (Table 3). However the total density of all trees and shrubs found in the year 2009 were 1114 and 1014 stems per hectare for the fenced and unfenced permanent plots respectively ( Table 3). The overall trees and seedling density that found in the fenced and unfenced was 52375 and 18410 stems per hectare in 2016 (Table   3). However, the total density of the permanent plots that found in the fenced and unfenced was 36014 and 25114 stems per hectare in 2009 (Table 3). Only two species of trees and shrubs were found to be dominating the permanent plots. The dominant species of the area relatively to the other 26 species based on relative density are Olea europaea sub species cuspidata and Juniperus procera which shared almost 88% of the total density.

Population Structure
The comprehensive patterns of the population structure of the whole woody species expressed in diameter class distribution of the fenced in 2016 ( Figure 6) showed an inverted J-shape. However, the diameter class distribution of the whole woody species in the unfenced block was an irregular shape ( Figure 6). More numbers of individuals are shown in the lower diameter and height classes of the fenced and unfenced permanent plot (Figures 6 & 7). However, higher regeneration potential was observed in the fenced.

Bonsai Growth
Olea europaea subspecies cuspidata is the highly and frequently browsed tree species by wild and domestic animals. The study is focused on this species on growth that means producing of new developing shoot that is proactive from the existing browsed species after fencing in terms of diameter and height. The average height of the bonsai growth was found to be 1.15m for the fenced.
The difference in bonsai height and diameter growth is highly significant (P<0.001). Fencing had significant contribution to bonsai diameter growth (P<0.001) with mean relative diameter growth of 10.5 ± 0.9cm ( Table 7). The total density of the bonsai that found in the fenced permanent plots were 343 and 434 is found in the unfenced permanent plots (Figure 7). Fencing has not brought a significant difference in total density of the browsed persistent seedling (bonsai) (Figure 7). Where; DSH is the relative diameter at sapling growth in cm, height is the relative height growth of the new leading shoot from the bonsai, significant difference among the management was tested (P < 0.05).

Net Change In Above and Below Ground Carbon Stocks:
The net total live tree carbon stock change of the fenced and unfenced block was found 3.56 Mg ha -1 and 0.9 Mg ha -1 respectively within eight years. That means the fenced management is increased its carbon stock by 0.44 Mg C ha -1 yr -1 and the unfenced permanent plot is also increased its carbon stock change by 0.11 Mg ha -1 yr -1 which is the average net gain annually.

Abundance, Density, Basal Area and Important Value Index:
Olea europaea subspecies cuspidata and Juniperus procera were the two most important dominant woody species found with high value of relative frequency, relative density and relative basal areas in the study area. They also constituted 88% of the total density. Aynekulu et al. [48] [50] where the stem density in the disturbed and undisturbed forests was not significantly different (P = 0.378). In line with the study of Senn et al. [51] the density of trees is higher in both fenced and unfenced areas.

Population Structure
This study indicated that the natural regeneration status of Juniperus procera in Desa"a forest is weak and in danger. This finding is supported by Aynekulu [48,29] and Aynekulu et al. [48] the regeneration of the native tree species in the open environment is very low and in a serious problem in the dry Afromontane forests. However, the regeneration of such forests were improved after exclosesing and protected from the interference of humans and animals [18,28,27,48]. The regeneration status of Juniperus procera is listed in IUCN Farjon [53] as in a red list of threatened species which is confirmed in this study that its regeneration is found to be very limited in both fenced and open environment of the dry Afromontane forests of Ethiopia that threatens its future existence [28,27,48].

Bonsai Growth
The bonsai height was significantly higher in the fenced than the unfenced. This is in agreement with the study by Aerts et al.
[54] and Giday [27]  Ground Carbon Stock over the unfenced one over the time. In line with this a study by the forest structure and tree density is higher and significant difference in the protected than the disturbed areas the above result of the forest population structure and density of trees is also affects the biomass and carbon stock of the forests. This is in agreement with the study by Yohannes et al. [56] higher total means carbon stock of above ground biomass was recorded in the least disturbed than the medium and higher disturbed forests. This is may be due to the trees and shrubs density is increased over a time has a significant effect (P = 0.02). However, on the unfenced over a time has not significant difference over a time. Because the tree density is increased over a time in both fenced and unfenced blocks.
This is may be due to in consistency with the above result of trees and shrubs stem density is not significantly differed (P = 0.373). As the trees and shrubs stem density per hectare is not that much changed over time. Because the biomass and carbon of a forest is expressed by the factor of diameter at breast height, height and wood density Chave et al. [43] affects the amount of carbon stored in the forest. This is similar with the study of Mokria et al.
[57] were found 19.3 ± 3.9 Mg C ha -1 at this forest area. This may be due to the trees and shrubs stages that are explained by the factor of diameter and height are high resistant to the disturbances that do not much affect by animals browsing and trampling because once they grow the animals cannot reach the leave of the trees and they may also develop their resistivity to the climate variability like moisture stress if once move to the stage of tree from the stage of seedling. According to the report USAID [58] showed that, long term fencing or protecting of an area to exclude grazing and browsing animals is higher carbon stock in the below ground and root as woody biomass and carbon increases. This is due to the minimizing of the disturbances by animals and humans.
Net Carbon Stock: This is in agreement with the study Yohannes et al. [56][57][58][59] conducted in West Shewa Oromia region Gedo forest, which described as forest disturbance increased, carbon stock in the above ground carbons, below ground carbons, litter, grasses and herbs carbons, soil carbon and total carbon stock density pools decrease. Higher total means for above ground carbon, below ground carbon, grasses, herbs and litters biomass carbon and total carbon density was obtained on the protected and or least disturbed stand [60][61][62][63][64][65][66][67][68].

Conclusion
The population status of tree species in the fenced permanent plot showed good regeneration progress. This indicates that though fencing the whole forest is practically difficult, exclosing and avoiding disturbances can among the indigenous species, Olive tree was found to have higher regeneration and IVI in the fenced permanent plot, hinting the possibility of restoring this species through exclosion of disturbances [69][70][71][72][73][74]. However, though Juniperus procera, the most dominant species in Desa"a, brought the highest IVI, it is in a danger with a very limited regeneration.

b.
Bonsai growth of the species Olea europaea subspecies cuspidata and developing new leading shoot from its persistent seedling has a significant conservation values and the most important as succession facilitator. So, protection of the staid potential of the forest areas of the dry Afromontane forest by different management to enhance the regeneration status and development of leading shoot from the persistent browsed Olea europaea subspecies cuspidata seedling is the best option to conserve the degraded dry evergreen Afromontane forests than recruitment from the seed bank.

c.
Future research should focus on over all regeneration potentials including vegetative propagation and soil seed bank regeneration capacity of the forest especially Juniperus procera, and total carbon stock including soil carbon of the permanent plot study area.
More research is needed to understand the relationship between species composition and above ground biomass carbon stock in the Desa"a dry forest.