Ethiopian church forests : opportunities and challenges for restoration

In Northern Ethiopia almost all dry Afromontane forests have been converted to open agricultural lands. Only small isolated fragments remain around churches ("church forests"), but these are many. This study analyses forest community structure and composition of the church forests, investigates major bottlenecks for regeneration of woody species, and explores opportunities and challenges for restoration. In this thesis, the following major questions are addressed: 1. How do forest structure, species composition and biodiversity vary<br/> across church forests and what are the major factors driving such variations?<br/> Factors considered are altitude, forest size and human influence. 2. What are the major bottlenecks in the regeneration of woody plants in<br/> church forests? These bottlenecks are studied for one to seven church forests and a<br/> major focus is given to the effects of soil seed bank and post-dispersal seed<br/> predation on seedling establishment, and of livestock grazing, microsite gradients<br/> and management interventions on seedling establishment, seedling survival and<br/> growth. The species and structural composition of 28 forests located at different altitudes (range 1816 to 3111 masl) and of various sizes (range 1.6 to 100 ha) was assessed, in relation to altitude, forest area, livestock grazing intensity and wood harvest intensity. A total of 168 woody species (100 tree species, 51 shrubs and 17 lianas) representing 69 families were recorded in the 28 church forests studied. Out of these species, 160 were indigenous and only eight were exotic (2 shrub and 6 tree species) representing 6 families. These forests accommodate many species represented by single individuals (rare species) and also many species found only in single plots (unique), which makes church forests priority of conservation efforts. Forests differed strongly in species number (15 to 78), basal area (4.8 to 111.5 m²/ha), density (&gt; 5 cm dbh: 267 to 1553/ha; &gt;cm diameter: 619 to 2421/ha and; seedlings: o to 5263/ha). Altitude is the main factor determining species composition of these forests. Our results showed that altitudinal gradient is the main determinant of differences in species composition among these forests. Maximum similarity in species composition was found with minimum altitude difference between forests. Geographical distance had only a weak effect on similarity. Therefore their vas altitudinal distribution gives these forests the opportunity to hold most of the biodiversity resources of the area. This can be confirmed by the number of specie: we found in our 28 church forests (168) which is more than the expected number 0 species (125 woody species) listed in the study area by South Gondar Zona Department of Agriculture. Structural composition of these forests, on the other hand, is determined by human influence. In the understorey the interaction effect of altitude with human influence (particularly cattle interference) determines the species composition. Not only the species composition but also the number of seedlings were severely affected by cattle interference and few, if any, seedlings were found in some of the forests. In the long run this would probably lead to a human effect on overstorey composition of these forests as well. Forest area did not show a significant effect. This implies that although large sized forests are a necessary element of successful reproduction of woody species, small patches and appropriate matrix management could be useful complements for biodiversity conservation. For many tree species regeneration in most of these forests is very limited. Possible reasons for this limitation are addressed in field experiments. Soil seed bank analysis on seven of these forests showed that these forests accumulate large quantities of persistent seeds of herbaceous species in the soil, but only five (6%) of the 91 woody species recorded in the standing vegetation of the seven forests were represented in their soil seed banks. Most of the tree species do not accumulate seeds in the soil. In order to investigate the long term behavior of seeds in forest soil, we assessed seed viability of five tree species in four sites of one forest after being buried 6, 12, or 18 months. Seed viability decreased sharply with burial time in soil for all species except for seeds of <i>Juniperus, </i>which still had 91% of viability after 18 months. Species significantly affect "the viability of the seeds after 18 months of burial. The quick decline of seed viability in forest soil indicates that the study species do not have a persistent soil seed bank. The fact that most of the dominant tree species do not accumulate seeds and maintain viability in the soil suggests that their regeneration from seeds would be prevented by removal of mature individuals in the standing vegetation. The experiment of post-dispersal seed predation on seeds of six tree species in one forest showed that 92% of the seeds were predated within 3.5 months. Therefore lack of persistent soil seed bank aggravated by intense seed predation undermines seed availability for regeneration in church forests. For four selected tree species in two forests this study showed that livestock grazing is one of the main bottlenecks hampering seedling establishment, seedling survival and seedling growth. Almost none of the sown seeds were able to germinate in unfenced plots (in unfenced plots germination was 4 and 5%, compared to 57.5 and 63.3% in fenced plots, data for Dengolt and Gelawdios respectively). In the fenced plots, seedling survival was higher (Dengolt = 65 and Gelawdios = 56%) and seedlings grew faster while there was no survival in unfenced plots. This implies that controlling livestock grazing is of paramount importance for both the internal regeneration of church forests and for restoration of the degraded surroundings. The study also explored how tree regeneration varies along the gradient from the forest interior to the edge and open fields, and differs between canopy gaps and closed canopy sites inside the forest. Seedling establishment was more successful inside the forest and in particular in the gaps within the forest. The exposure effect between forest interior and exterior was not analogous to the exposure effect of gaps inside the forest. This result suggests that germination may be primarily influenced by moisture availability along the forest interior-open field gradient. Though seedling establishment is higher inside the forest, seedlings grew more rapidly and survived better on the outer edge of the forest. Our result revealed that the negative effect of edge on regeneration is seen on the inner edge while the outer edge has positive effect on seedlings survival. As a result of the interplay effects of light and moisture, the outer forest edge might become the optimal place for seedling survival and growth. This may indicate that seedlings can colonize and restore the forests on the degraded land from immediate surrounding areas of the church forests, given that the surrounding land is protected from grazing intervention and farming. Within the forest, light in gaps favored survival and growth of seedlings both in dry and wet periods. Perhaps, the soil moisture in gaps inside the forest was not as depleted as that outside the forest in the dry period. In general, the quantitative effect of microsite differed with species. Restoration experiments using management interventions (sowing seeds, planting seedlings, weeding, litter removal, soil scarification) showed that the combined effect of seed sowing and litter removal increased seedling establishment significantly for all the species. This implies that insufficient availability of seeds could be one of the constraints for regeneration in church forests. Weeding did not improve seedling survival and growth of the species, and had even a negative effect on some of them. The future existence of the woody flora and vegetation characteristic of dry Afromontane areas in Ethiopia depends on effective conservation and sustainable utilization of the remnant natural forest patches. If not possible to conserve all remnant forests, selection of church forests across altitudinal variation is important to accommodate most of the species diversity of the area. Maintaining viable populations in the forests and providing connections between forests is pivotal in this respect. Excluding cattle interference and reducing intensity of wood harvest are a prerequisite to facilitate regeneration in church forests. To offset the seed limitation sowing seeds combined with litter removal and slight soil scarification can be very useful. Sowing of seeds should be integrated with active reduction of predation and herbivory (e.g. in the form of enclosures and providing alternative food sources for predators). Seedling transplanting is another alternative to overcome some of the bottlenecks of germination especially moisture deficit and intense seed predation, but weeding might not be needed. Within the forest, opening up of the canopy may be required to facilitate seed germination and seedling performance depending on the species. Along the gradient from forest interior to edge and open field, seed sowing will be effective in the interior microsite, while transplanting seedling in the outer forest edge gives better survival and growth. These measures improve tree regeneration. Interconnecting these remnant forests by vegetation corridors following natural terrain or stream lines, or reducing the distance between them by creating buffer areas and plantations around them, and developing more patches in the landscapes are possible management activities. These will facilitate propagule and germplasm flow and ultimately may sustain these forests and help restoring the surrounding landscape.