Quantifying threats to biodiversity from invasive alien plants and other factors: a case study from the Cape Floristic Region

Introduction It is often stated that biological invasions pose the second most pressing threat to biodiversity after direct habitat transformation. Yet this assertion, which has crucial policy implications for conservation, land-use planning and restoration, has yet to be quantitatively tested, except on the basis of analysis of threatening factors listed for rare species. A more detailed assessment of the spatial dynamics of biological invasions and other threat factors is needed to inform effective conservation management. South Africa’s Cape Floristic Region offers opportunities to advance our understanding in this area. First, the area is spectacularly rich in plant species, very detailed distribution data are available for some key taxa, and there is excellent information on environmental factors, a key prerequisite for spatial modelling. Second, the extent of the main types of habitat transformation has been mapped. Although the vulnerability of remaining natural areas to transformation has been assessed, less attention has been given to characterizing how different types of transformation, including agriculture, forestry, urbanization and invasive alien plants, have affected features of natural communities, including species richness and rare species representation. This paper presents a novel method for producing an essential prerequisite for such an analysis, a prediction of species distributions and species richness in the absence of transformation. Unlike standard approaches to quantifying past and potential effects of alien plant invasions, the method presented here, using the output from Bayesian hierarchical regression models together with maps of transformation derived primarily from satellite imagery, facilitates a description of the effects of different kinds of transformation on individual species distributions and species richness. One aim of this study was to assist in objective priority formulation for management programmes aimed at clearing invasive alien plants in natural and seminatural vegetation. This study used spatially explicit, hierarchical regression models to predict the distributions of a representative set of species in the protea family (Proteaceae), one of the characteristic families of fynbos vegetation in the Cape Floristic Region. The models computed a ‘potential range’ for each species in the form of a probability surface that represents predicted probability of occurrence in the absence of habitat transformation. This potential range was then adjusted by the percentage of each grid cell that has been transformed to obtain a ‘transformed range’. Unlike previous efforts to model species distributions and richness, the distributions of all the species were modelled simultaneously, enabling predictions for each species to be informed by a common spatial effect. By summing across species, a prediction of potential species richness was obtained. Using the maps of habitat transformation prepared for the recent Cape Action Plan for the Environment, we compared the effects of alien plant invasions with those of agriculture, forestry, and urbanization. We related the distributions of these impacts to patterns of potential richness and to individual species distributions to address three questions: 1) Where do the various types of habitat transformation occur in relation to areas of high species richness? 2) What have been the consequences of these changes so far on richness (using Proteaceae as surrogates) and on rare and common species? 3) What types of transformation present the most immediate threats to diversity and rare species?