Spatial and temporal variation in species-area relationships in the Fynbos biological hotspot


  • Lisa L. Manne,

  • Paul H. Williams,

  • Guy F. Midgley,

  • Wilfried Thuiller,

  • Tony Rebelo,

  • Lee Hannah

L. Manne (, Biological Sciences, Univ. of Toronto at Scarborough, Scarborough, ON MIC 1A4, Canada. – P. H. Williams, Natural History Museum of London, Cromwell Road, London SW7 5BD, U.K. – G. F. Midgley and T. Rebelo, South African National Biodiversity Inst. P/Bag X7, Claremont 7735, Cape Town, South Africa. – W. Thuiller, Laboratoire d'Ecologie Alpine, UMR CNRS 5553, Univ. J. Fourier, BP 53, FR-38041 Grenoble Cedex 9, France. – L. Hannah, Conservation International, 1919 M Street NW, Suite 600, Washington, DC 20036, USA.


Species-area relations (SARs) are among the few recognized general patterns of ecology, are empirical relations giving the number of species found within an area of a given size and were initially formulated for island environments. The use of SARs has been extended to mainland environments, and to give baseline estimates of extinction rates attending habitat loss. Using current species distributions based on atlas data, we examined the spatial variation of rates of species accumulation and species-area curves for Proteaceae species for all one-minute by one-minute areas within the Cape Floristic Region, South Africa. We compared SARs for current distributions to those generated from modeled future Protea distributions following climate change. Within one biome and for two different scales, there exists a very large spatial variation in turnover rates for current Proteaceae distributions, and we show that these rates will not remain constant as climate warming progresses. As climate changes in coming years, some areas will gain species due to migration, as other areas lose species, and still other areas maintain current rates of species accumulation/turnover. Both current and future distributions show highly variable rates of species accumulation across the landscape. This means that an average species-area relationship will hide a very large interval of variation among SARs, for both current and future Proteaceae distributions. The use of species-area relations to estimate species extinctions following loss of current habitat, or loss of future climatically-suitable area is likely to result in erroneous predictions.