The impact of forest on pest damage, pollinators and pollination services in an Ethiopian agricultural landscape

Abstract: The distribution of wild biodiversity in agroecosystems affect crop performance and yield in various ways. In this thesis I have studied the impact of wild biodiversity, in terms of trees and forest structures, on crop pests, pollinators and the pollination services provided in a heterogeneous landscape in southwestern Ethiopia. Coffee, Coffea arabica, is a forest shrub native to Ethiopia and is grown in most wooded areas in the landscape where I conducted my studies. Wild coffee is still found in remote parts of the forests in the landscape. For my first paper, I surveyed pest damage on coffee in coffee forest sites, where some sites were situated in continuous forest and some in isolated forest patches. I found the variation in pest damage frequency to mainly be among coffee plants within a site, rather than among sites, which indicates the importance of local processes. However, some pests were clearly connected to the forest habitat, such as the olive baboon.In my second study, I surveyed pollinators visiting coffee flowers across a gradient of shade-tree structures. I found the semi-wild honeybee to be the dominating flower visitor. The abundance of the honeybee was not related to shade-tree structures, but to amount of coffee flower resources in the site. On the other hand, other pollinators, which included other bee species and hoverflies, were positively affected by more shade trees in the site.In my third study I investigated how the forest cover affected local bee communities in the agricultural landscape. Moreover, I investigated if this relationship differed between the dry and rainy season. The distribution of food resources for bees changes between the seasons, which may affect the bees. Most trees, fruit trees and coffee, which are patchy resources, flowers in the dry season, whereas most herbs and annual crops, which are more evenly spread resources, flowers during the rainy season. I found a clear turnover in bee species composition between the dry and rainy season, with more mobile species in the dry season. Increased forest cover in the surrounding landscape had a positive impact on bee abundance and species richness. However, the impact did not change between seasons.In my fourth study I evaluated the pollination success and pollen limitation of a common oil crop in the landscape in relation to forest cover. I found severe pollen limitation across the landscape, which may be related to the observed low bee abundances. The pollen limitation was not related to surrounding forest cover.In conclusion, I have found the forest and wooded habitats to impact several mobile animals and pathogens in our study landscape, which in turn affect people. However, there is large complexity in nature and general relationships between forest structures and all crop related organisms may be unlikely to find. Various species are dependent on different resources, at different spatial scales and are interacting with several other species. To develop management strategies for increased pollination services, for reduced pest damage or for conservation in the landscape, more species-specific knowledge is needed.