Assessing the Impacts of Future Climate Change on Protected Area Networks: A Method to Simulate Individual Species��� Responses Stephen G. Willis �� Dave G. Hole �� Yvonne C. Collingham �� Geoff Hilton �� Carsten Rahbek �� Brian Huntley Received: 5 July 2006 / Accepted: 28 March 2008 / Published online: 20 May 2008 �� Springer Science+Business Media, LLC 2008 Abstract Global climate change, along with continued habitat loss and fragmentation, is now recognized as being a major threat to future biodiversity. There is a very real threat to species, arising from the need to shift their ranges in the future to track regions of suitable climate. The Important Bird Area (IBA) network is a series of sites designed to conserve avian diversity in the face of current threats from factors such as habitat loss and fragmentation. However, in common with other networks, the IBA net- work is based on the assumption that the climate will remain unchanged in the future. In this article, we provide a method to simulate the occurrence of species of conser- vation concern in protected areas, which could be used as a first-step approach to assess the potential impacts of cli- mate change upon such species in protected areas. We use species-climate response surface models to relate the occurrence of 12 biome-restricted African species to cli- mate data at a coarse (quarter degree-degree latitude- longitude) resolution and then intersect the grid model output with IBA outlines to simulate the occurrence of the species in South African IBAs. Our results demonstrate that this relatively simple technique provides good simu- lations of current species��� occurrence in protected areas. We then use basic habitat data for IBAs along with habitat preference data for the species to reduce over-prediction and further improve predictive ability. This approach can be used with future climate change scenarios to highlight vulnerable species in IBAs in the future and allow practical recommendations to be made to enhance the IBA network and minimize the predicted impacts of climate change. Keywords Biodiversity Climate change Climate envelope modeling Protected area network Introduction Over recent decades, species have been declining and becoming extinct, both locally and globally, at an alarming rate (BirdLife-International 2000 Brooks and others 2002 Pimm and others 1995 Thomas and others 2004b). The conservation of biodiversity is therefore an urgent priority and much effort has been put into identifying localities of maximum diversity (Balmford and others 2001 De Klerk and others 2002 Myers and others 2000) and the protection of such areas (De Klerk and others 2004 Muriuki and others 1997). BirdLife International���s Important Bird Areas (IBA) network is a biodiversity conservation network that iden- tifies priority sites for the protection of the global avifauna. As with most conservation networks, algorithms for the selection of priority sites are based entirely upon current species��� distributions. Such networks are essential for the urgent protection of species of high priority due to their threat status, restricted distributions, or other factors. However, they make no allowance for the fact that species��� ranges are dynamic over time, and thus, alone, these net- works may be inadequate for the conservation of species S. G. Willis (&) D. G. Hole Y. C. Collingham B. Huntley Institute of Ecosystem Science, School of Biological and Biomedical Sciences, University of Durham, Durham DH1 3LE, UK e-mail: s.g.willis@dur.ac.uk G. Hilton Royal Society for the Protection of Birds, The Lodge, Sandy, Beds SG19 2DL, UK C. Rahbek Institute of Biology, University of Copenhagen, Universitetsparken 15, 2100 Copenhagen, Denmark 123 Environmental Management (2009) 43:836���845 DOI 10.1007/s00267-008-9125-3

with shifting ranges. Consequently, there is an equally urgent need to identify those sites that will remain impor- tant for species over the medium to long term. IBAs predicted to remain resilient to change in the future should be prioritized for protection, and the existing network should be enhanced to minimize the predicted impacts of climate change. Global climate change is one major factor that has already caused recent alterations in species��� ranges (Parmesan and others 1999 Parmesan and Yohe 2003 Root and others 2003 Walther and others 2002 Warren and others 2001) and is likely to become increasingly important in altering species��� ranges over the current cen- tury (Hill and others 2002 Thomas and others 2004a). Biodiversity in Africa Tropical and equatorial ecosystems are noted for their relatively high biodiversity tropical latitudes also contain a high proportion of the world���s ������biodiversity hotspots,������ characterized by their concentrations of endemic species (Myers and others 2000). Of the top 25 global hotspots recognized by Myers and others (2000), four lie in main- land sub-Saharan Africa, while Madagascar represents a fifth. Sub-Saharan Africa plays host to over 1900 bird species, approximately 20% of species globally. Of these, 947 are confined solely to the Afrotropical region, around 340 being of global conservation concern (Fishpool and Evans 2001). In addition to the hotspots of endemism, sub- Saharan Africa has many other areas that harbor local concentrations of biodiversity. The mountainous areas of equatorial east Africa are one such example, with high- altitude cloud forests characterized by a suite of ecologi- cally restricted species found only in these spatially limited habitats. During recent decades, similar cloud forests elsewhere have suffered biodiversity losses, which corre- spond to regional warming and reduced incidence of cloud formation in the forests, especially during the dry season (Pounds and others 1999). Comparisons of future climate scenarios and current climate have revealed that in the future there is likely to be a substantial loss of climate types in mountain regions of Africa and an increase in novel, no-analogue climates (Williams and others 2007). Predicted Climate Change Global climate is expected to change substantially in the course of the present century, primarily as a consequence of anthropogenic greenhouse gas emissions (Christensen and others 2007). In many regions at tropical and sub- tropical latitudes, the most important changes are likely to be in precipitation patterns, both spatially and temporally. The recent IPCC Fourth Assessment Report projects that by 2080���99 declines in rainfall will be likely in most subtropical areas, with increases projected in areas of regional tropical precipitation maxima. Intensity of pre- cipitation is also projected to increase, particularly in tropical areas of increasing precipitation, with a tendency for drying of mid-continental areas (Meehl and others 2007). In the case of the African continent, there is a general consensus among GCM simulations indicating modest increases in mean annual precipitation over equa- torial Africa, particularly eastern regions. Conversely, northern, southern and parts of western Africa are projected to have reduced precipitation in most models, soil moisture being markedly reduced as a result (Meehl and others 2007). Increases in mean annual precipitation in the equatorial latitudes, however, often mask seasonal shifts. In equatorial east Africa, for example, GCM simulations indicate modest changes in precipitation during June��� August but substantial increases in precipitation during December���February (Meehl and others 2007). Tempera- ture increases projected for Africa by 2080���99, while modest compared to high latitudes, are nonetheless 4���5��C higher than temperatures recorded in 1980���99, based on 20 general circulation models and a medium-high emissions scenario (Boko and others 2007), with less warming in equatorial and coastal areas. Predicting Ranges Using Climate Data Large-scale distribution patterns of many species, from autotrophs through to top predators have been reliably simulated using climate variables (H-Acevedo and Currie 2003 Huntley and others 2004). The geographical ranges of a majority of the bird species of Europe and Africa can be been modeled well using a limited number of climatic variables (Huntley and others 2006, 2007). There is evi- dence, both from the Quaternary palaeoecological record (Graham and Grimm 1990 Huntley and others 1997) and from recent observational studies (Parmesan and others 1999 Parmesan and Yohe 2003 Root and others 2003 Walther and others 2002 Warren and others 2001), which shows species responding to environmental changes pri- marily by shifting their geographical ranges. Simulations of various European plant and animal species��� potential future distributions reveal 21st century climate changes may shift species��� potential range boundaries [1000 km (Hill and others 2002 Huntley and others 1995 Huntley and others 2008). There is an urgent need to investigate the extent to which ranges of African species might also shift, and to modify conservation management strategies accordingly. Environmental Management (2009) 43:836���845 837 123