Disturbance, diversity and distributions in Central African rain forest

The aim of this study is to gain insight in the impact of human land use on plant community composition, diversity and levels of endemism in Central African rain forest. Human disturbance in this region is causing large-scale habitat degradation. The two most widespread forms of land use are selective logging and shifting cultivation. Assessment of the long-term effects of these land uses on plant species composition will provide elements for the identification of effective conservation measures and sustainable forms of forest use.Disturbances are relatively discrete events in time that cause high mortality of biomass and change the structure of populations, communities or ecosystems. Individual plants and species differ in their ability to claim the previously utilised space and resources, and therefore disturbance events may cause (temporary) shifts in species composition and diversity. In large parts of the African rain forest biome, the most important natural disturbance regime is gap-phase dynamics in which relatively small canopy openings are made by falling branches or trees. Larger-scale disturbances such as landslides, volcanic activity and large-scale river dynamics, have been extremely rare during the last millennia. Therefore, the present set of species is likely to be adapted to disturbance regimes characterised by frequent small-scale disturbances. Analogous to gap-phase dynamics, human land use can be considered as a disturbance. However, human induced disturbances are generally larger, more frequent and more severe. As a result, human land use may have long-term impacts on plant species composition and diversity.The fieldwork for this thesis was conducted in the main research site of the Tropenbos-Cameroon Programme; an area of 2000 km 2 of forestland in South Cameroon Vegetation was sampled in old growth forest, logging gaps of 5, 14 and 27 years old, and in abandoned shifting cultivation fields of 10-20, 30-40 and 50-60 years old. Within plots, all terrestrial vascular plants (including all growth forms) were recorded. In total 10.1 ha was surveyed allowing the detailed analysis of human impact on full plant species composition. Much effort was devoted to plant identification and in total some 11 000 herbarium vouchers were collected, processed and sent to specialists for identification. Voucher material was stored in theNationaalHerbarium Nederland - WageningenUniversity branch and in the IRAD/TropenbosField Herbarium inKribi. Some 75% of the plants in the survey identified to species level and an additional 20% was systematically categorised asmorphospecies. In total 1264 species were identified to species level. These included 261 species with ranges restricted to Lower Guinea (South Nigeria - Gabon) of which 51 are confined to the rain forest region of Cameroon.Tropical rain forests are often regarded as being undisturbed by humans. In Chapter 2, we analyse the disturbance history of 16 ha of structurally complex and species-rich 'old growth' rain forest (data collected by forestry research project of theTropenbos-Cameroon Programme). The recruitment preference of tree species along a disturbance gradient, ranging from shifting cultivation fields, to canopy gaps and old growth forest, was compared to present-day tree species composition. In nine plots out of 16, older (larger) trees preferred shifting cultivation fields for recruitment while younger trees recruited in small canopy gaps and under closed canopy. These results indicate that these plots once experienced a disturbance regime that included larger-scale disturbances. Combined, the pattern of disturbed and undisturbed plots, the high frequency of charcoal in the forest soil and anthropological data strongly suggest that humans caused these disturbances. The estimated date of these disturbances is 300-400 years ago. Surprisingly, species richness at larger scales was found to be lower in historically disturbed sites compared to undisturbed sites. Therefore, present-day species composition and diversity of old growth forests still reflects historical human impacts.Outside protected areas, forests are subject to logging and shifting cultivation and secondary forests are therefore becoming important in many Central African landscapes. Chapter 3 examines the potential of secondary vegetation to contribute to biodiversity conservation. The results indicate that vegetation recovery in logging plots and abandoned shifting cultivation fields is relatively quick, and in most aspects surprisingly complete. A notable exception is the poor recovery of endemics in shifting cultivation fields. We found that even after 60 years the proportion of endemic species was still significantly lower than in old growth forest. In light of the fast recovery of all other vegetation characteristics (including species richness and floristic composition), we conclude that secondary vegetation can contribute biodiversity conservation e.g. as buffer zones around protected areas.In Chapter 4, we analyse the relative importance of local and regional processes for structuring species composition during succession. Local processes refer to the ability of species to compete successfully with other species and avoid predation and pathogen attack. Regional processes refer to dispersal and colonisation. Both processes have been suggested to explain the typically high species richness in tropical rain forests. Our results indicate that local processes are especially important during the early stages of succession, whereas regional processes are especially important during the later stages. However, large differences were observed between different species groups. Regional processes mainly governed composition of large tree species (maximum height ≥ 15 m). A similar but smaller effect was observed in small tree species (3-15 m) and shrub species. Local processes structured composition of terrestrial herb species. In general, woody and non-woody climbers were widely distributed generalists with a very similar set of species occurring in all landscape mosaics and in all succession stages. The differences between species groups implies that effective conservation management requires insight in the importance of local and regional processes for the recruitment of target species (e.g. endemic species). A blanket conservation treatment for all species is unlikely to address adequately the specific sensitivity of species with high conservation value to habitat quality and habitat fragmentation.The general lack of information on biodiversity patterns is a serious problem for conservation planning in most tropical rain forest regions. With ongoing habitat destruction, conservation priorities must be identified quickly. Therefore optimal use should be made of all currently available sources of information. However, direct comparison of results is problematic if assessment methods differ. In Chapter 5, we make a first attempt to reconcile different assessments by taking into account their methodological differences. The key factors affecting the general shape and position of the species-area curve found through partial sampling are (1) the total extent in which observations are made, (2) the spatial distribution of the observations, (3) the proportion of the total extent sampled, (4) the proportion of the individuals in the sampled area that was included in the survey, and (5) the proportion of the included individuals that was successfully identified. Through simulations, the effects of partial sampling of these factors on observed species richness was identified. To test the method, we compared four botanical surveys conducted in the same area of lowland old growth rain forest. The surveys included were (1) reconnaissance scale vegetation survey, (2) detailed botanical assessment (100% individuals), (3) incomplete botanical assessment (10% individuals), and (4) herbarium collections. Correcting for partial sampling and scaling the results to extent greatly increased the comparability between assessments. This first attempt to reconcile methodologically different surveys suggests that species -area relations can be reconstructed from incomplete sample data if the key characteristics of the methods can be statistically described. The study provides an outline for optimising the use of existing datasets in the evaluation of conservation needs in tropical rain forest areas.In Chapter 6, I present an overview of the main effects of human land use on plant diversity in Central African forest. Rain forests are highly dynamic on all spatial and temporal scales. Present-day species composition and diversity reflects their cumulativebiogeographicalhistory. Therefore both present-day and historical disturbance regimes should be considered to understand current patterns of diversity and to predict its responses to future disturbances. The biotic andabioticprocesses that influence diversity vary with the scale of organisation of biological and ecological systems (i.e. community, ecosystem, landscape,region). While human land use obviously disrupts local communities, the impact of humans is also evident on much larger scales. As a result of large-scale forest degradation, fragmentation and global climate change, species composition of most Central African rain forests is likely to undergo changes in the near future. Conservation management should aim at increasing the survival chances of endemic species and species with poor dispersal capacity. In light of the present ecological insights and the uncertainty of the upcoming climate changes, it seems wise to invest in large networks of protected areas. Moreover, it is unlikely that areas managed for timber production will be beneficial for the conservation of characteristic plant diversity unless damage control is rigid. The expected increase in seasonality in large parts of Central Africa, combined with forest fragmentation and canopy opening, is likely to increase the abundance of pioneer species while species of concern to conservation are likely to decline. To increase the effectiveness of conservation management, insight is required in the mechanisms that make species and forest systems vulnerable to human induced disturbances, including global climate change.