Modified forests are vital for species communities and ecological functionality in a heterogeneous South African landscape

Land-use change is a major threat to forest ecosystems worldwide. Therefore, understanding the effects of human forest modification on biodiversity is an important task for conservation ecologists. The main objective of my dissertation was to evaluate how different intensities of forest modification contribute to the maintenance of species diversity and ecosystem functionality in a human-modified landscape. For this purpose, I based my studies in a heterogeneous landscape around two nature reserves, Vernon Crookes and Oribi Gorge, in KwaZulu-Natal, South Africa. I selected six most representative types of scarp forest modification – ranging from continuous forest and natural forest fragments in nature reserves to fragments within plantations and agricultural matrix, forested gardens, and secondary forest. In a total of 36 study sites, I assessed flower-visiting insects using insect traps and recorded local bird assemblages with point counts. Further, I observed flower visitation and seed removal on the native and widespread tree Celtis africana (Ulmacea) to analyse whether forest modification affects pollination and seed dispersal services. To assess how forest configuration affects the dispersal of animals, I carried out an in-depth study on the movement behaviour of bird assemblages within and among forest patches in the Vernon Crookes region. By means of direct observations and bird mist-netting, I followed up bird movements across nine forest fragments belonging to three different forest types. In all these three projects I give special attention to the responses of the different functional groups of a species community. The richness of flower-visiting insects, community composition and flower visitation on C. africana differed significantly among the different forest types and between two study seasons in 2009 and 2010. Both flower visitor richness and flower visitation rates were strongly enhanced in the human-modified forests. This could be explained by a high abundance of large-bodied pollinators in these sites. In particular, feral honey bees (Apis mellifera) played a major role in the pollination of C. africana trees located in forest fragments within plantations and agriculture, forested gardens and secondary forests. However, effective fruit set of C. africana was not enhanced by an increase of flower visitation, possibly due to the tree’s capability of wind pollination. This implies that even though forest modification can strongly alter insect assemblages, pollination services for trees with unspecialized flowers may remain resilient at a landscape scale. Bird species richness was not significantly different among forest types. However, I found a significant increase in bird abundance in modified forests. In particular, fragments within agriculture, forested gardens, and secondary forests attracted a large number of forest generalists, shrubland and open country species. The abundance of forest specialists however, was much lower in modified forests. Changes in the composition of bird functional groups were also confirmed by multivariate analysis, which clearly separated bird communities by forest type. I found the highest abundance of frugivores visiting C. africana in natural forest fragments, fragments within agriculture, forested gardens and secondary forests. That was also true for the estimated total fruit removal per C. africana tree, even though the differences among the forest types were not significant. In summary, I could show that overall bird abundance and seed removal services can be enhanced in modified forests. However, the results also underline the importance of protected natural forest for bird specialist species sensitive to human disturbance. I found a very high movement activity of the overall bird community among the nine forest fragments that was significantly structured by bird functional groups. Especially, frugivorous birds, forest specialists and large-bodied species showed the highest dispersal abilities across the landscape. These results might be facilitated by overall high fragment quality, providing food and shelter, as well as the close proximity among the forest fragments within the landscape. Yet, a fourth-corner analysis revealed that even though modified forests were rather attractive to frugivores, forest specialists as well as large-bodied species, there was still a high affinity of the latter functional groups to natural forest fragments, close canopy cover and large fragment size. Only a small proportion of the overall bird community was recorded to steadily persist in the forest fragments. In particular, patches in the agricultural landscape were frequently used by resident insectivores and forest generalists. Ultimately, these findings suggest that remnant forest fragments may represent valuable stepping-stones as well as permanent habitat for many forest birds and thus, will help to maintain regional bird assemblages in human-modified landscapes. Overall, my results strongly suggest that modified forests contribute to the maintenance of species diversity and ecosystem functionality in a human-modified landscape. With respect to a vast increase of human-modified forests worldwide, evidence of a high conservation potential of these habitats is encouraging news for conservation managers. In particular, modified forests that are located in close proximity to protected areas have high conservation priority as they may expand buffer zones around natural forests in human-modified landscapes. Generalizations, however, should be considered with caution. My findings strongly emphasize that human-modified forests do not completely compensate for the overall loss of natural habitat. High sensitivity of forest specialist species and overall changes in local community composition demonstrate that natural forests are essential to maintain species diversity at a larger scale. Additionally, high flexibility towards habitat changes of many species in the study region might be based on the patchy historic distribution of scarp forest that has strongly been determined by terrain and orographic conditions of the environment. Thus, it is possible that an evolutionary adaptation has lessened the vulnerability of the region’s fauna and flora to the present anthropogenic forest fragmentation. Ultimately, most of the forest types in the study region are characterized by high habitat quality, including for example resource availability, structural heterogeneity and close proximity to further forest patches, so that altogether, they contribute to the high species diversity. Consequently, the maintenance of structurally rich forest habitat is essential to maintain species diversity and ecological functionality in human-modified landscapes