Botswana Notes and Records, Volume 45 96 Estimating Biodiversity in Remote Areas, Using Existing Vegetation Data: The Ngamiland Region Cornelis Vanderpost * Susan Ringrose † and Mike Murray-Hudson ‡ Abstract In data-poor regions, especially when they are large and remote, the measurement of biodiversity presents considerable challenges. This paper explores a way of estimating regional patterns of biodiversity through a combination of land-cover feld mapping, remote sensing and interpretative GIS techniques. The results show spatial variations of potential biodiversity in the remote Ngamiland region of Botswana, with areas of higher variability of land-cover classes indicative of higher degrees of biodiversity. The methodology is potentially replicable in other data-poor regions in developing countries. Introduction Diversity is considered important in terms of the overall environmental stability of ecosystems. More complex systems are often more resilient, and have a greater potential to adapt to change in environmental parameters (Odum 1971). Biological diversity is often expressed hierarchically, in terms of species diversity at a particular site or within a given community, between communities or habitats, or as total species diversity in a region (Whittaker 1972 and Ricotta 2005). These indices can be related through additive or multiplicative partitioning of the smaller scale measures, as contributors to the larger scale measure (Ricotta 2005 and Crist 2003). The issue of biodiversity at different geographic scales presents various challenges as aspects of species or ecosystem diversity can be considered at various spatial scales as landscape or regional variety at, for example the eco-subregion or ecoregion level (Sorokine, Bittner & Renschler 2004). Ecosystem diversity, a concept that encompasses the variety of habitats occurring within a region or the mosaic of patches found within a landscape, is harder to measure than species or genetic diversity because the ‘boundaries’ of communities (associations of species) and ecosystems are potentially elusive (World Resources Institute, World Conservation Union and United Nations Environment Programme 1992). Whittaker, Willis and Field (2001) suggest therefore to use the terms local, landscape and macro-scale as a more intuitive framework for addressing issues of geographic scale. Even then, the measurement of diversity and its intrinsic meaning continue to present challenges to the scientifc community, particularly with respect to the interpretation of the various measures (Giles and Trani 1999). Nevertheless, politicians and scientists alike now agree that a priority list of global centres for preservation of biological diversity is required (Wood, Stedman-Edwards and Mang 2000). Hence, the need for biodiversity maps and other data. The measurement of species diversity over large regions in third world countries with relatively poor ground data presents a major challenge (Williams 1996). Beta-diversity, also referred to as between-habitat diversity and spatial species turnover or differentiation (Koleff, Gaston and Lennon 2003, Whittaker 1972), describes the increase in diversity or species numbers when sub- areas are combined. Its measurement typically requires multiple closely spaced transects with detailed species information covering the entire region, making this a prohibitively data intensive procedure for large regions. Although higher taxon richness (using genera or families) has been suggested as a * Cornelis Vanderpost, Okavango Research Institute, University of Botswana. Email: cvanderpost@outlook.com † Susan Ringrose, Okavango Research Institute, University of Botswana. Email: sringrose@orc.ub.bw ‡ Mike Murray-Hudson, Okavango Research Institute, University of Botswana. Email: mmurray-hudson@ori.ub.bw