Current patterns of habitat transformation and future threats to biodiversity in terrestrial ecosystems of the Cape Floristic Region, South Africa

The formulation of an effective strategic plan for biodiversity conservation in the Cape Floristic Region (CFR) requires an assessment of the current situation with regard to habitat transformation, and an explicit framework for predicting the likelihood of remaining habitat (i.e. that potentially available for conservation) being transformed. This paper presents the results of a detailed assessment of the current and future extent of three important factors that threaten biodiversity in the CFR: cultivation for intensive agriculture (including commercial forestry plantations), urbanisation, and stands of invasive (self-sown) alien trees and shrubs. The extent of habitat transformation was mapped at the scale of 1:250,000, using primarily satellite imagery. We compared models derived from a rule-based approach relying on expert knowledge and a regression-tree technique to identify other areas likely to be affected by these factors in future. Cultivation for agriculture has transformed 25.9% of the CFR and dense stands of woody alien plants and urban areas each cover 1.6%. Both models predict that at least 30% of the currently remaining natural vegetation could be transformed within 20 years. There was an overall accuracy of 73% between both models although significant differences were found for some habitat types. Spatial predictions of future agriculture threats derived from the rule-based approach were overestimated relative to the statistical approach, whereas future alien spread was underestimated. Threat assessment was used to derive conservation targets for subsequent stages of conservation planning for the CFR. The importance of integrating vulnerability knowledge into conservation planning is discussed. The choice of vulnerability analysis (future habitat degradation and/or impact on biological entities) and methods will depend on the complexity of the threatening processes and the availability of spatial data. (C) 2003 Elsevier Science Ltd. All rights reserved.