Complementarity of Species Distributions as a Tool for Prioritising Conservation Actions in Africa: Testing the Efficiency of Using Coarse-Scale Distribution Data

Predicting species distribution at range margins: testing the effects of study area extent, resolution and threshold selection in the Sahara–Sahel transition zone

“…In fact, continental models tend to be ambiguous at range margins, failing to detect isolated small patches or local and fragmented distributions (Fjelds a & Tushabe, 2005). Overall, continental models at coarse resolution (C10) showed a good level of agreement with species' occurrence data.…”

“…However, zooming at the edge of range (cC10), they tended to overestimate species distributions, as observed by the low correct classification rates of absence data in comparison with validation data. In fact, continental models tend to be ambiguous at range margins, failing to detect isolated small patches or local and fragmented distributions (Fjelds a & Tushabe, 2005). Models calibrated using coarser resolutions may substantially overestimate potentially suitable areas in comparison with those built with finer resolution data .…”

“…At range limits, the distribution of local populations may be restricted to patches of suboptimal habitats, which may present local characteristics that are distinct from the species' complete range (Braunisch et al, 2008). When local conditions represent a minor part of the species' environmental range, global models built with coarser resolutions may produce biased predictions at range margins (Braunisch et al, 2008), such as overestimating local species distributions and missing finer distributional details or local distribution gaps (Osborne & Su arez-Seoane, 2002;Fjelds a & Tushabe, 2005;Hernandez et al, 2006). When local conditions represent a minor part of the species' environmental range, global models built with coarser resolutions may produce biased predictions at range margins (Braunisch et al, 2008), such as overestimating local species distributions and missing finer distributional details or local distribution gaps (Osborne & Su arez-Seoane, 2002;Fjelds a & Tushabe, 2005;Hernandez et al, 2006).…”

“…As such, ENM predictions at range margins will largely depend on the environmental features at a small scale, because different factors may control range limits at different parts of the range. When local conditions represent a minor part of the species' environmental range, global models built with coarser resolutions may produce biased predictions at range margins (Braunisch et al, 2008), such as overestimating local species distributions and missing finer distributional details or local distribution gaps (Osborne & Su arez-Seoane, 2002;Fjelds a & Tushabe, 2005;Hernandez et al, 2006). These patterns could be even more pronounced in range margins located in transition zones between climatic or habitat extremes because the discriminating ability of ENMs may be affected by differences in habitat availability or selected by species (Osborne & Su arez-Seoane, 2002).…”

The partitioning of Africa: statistically defined biogeographical regions in sub‐Saharan Africa

“…The development of the databases has involved a review of all available scientific literature, consultations with more than 25 experts and visits to 12 museums in eight countries, as well as new fieldwork in the East African Region (e.g. Burgess et al, 1998aBurgess et al, , 2007bBrooks et al, 2001;de Klerk et al, 2002a,b;Fjeldså & Tushabe, 2005;Linder et al, 2005;Fjeldså et al, 2010; databases available at http://130.225.211.158/subsaharanafrica/subsaharan.htm, accessed 13 January 2011, Zoological Museum, University of Copenhagen).…”