Answers are grounded in this paper and cite their sources.
In the last two decades, several quantitative techniques for assessing conservation priorities have been developed, based on data about the distribution of species (Reid 1998, Williams 1998, Margules and Pressey 2000). These methods have been applied extensively in temperate regions such as North America at both the state (Csuti et al. 1997) and national (Dobson et al. 1997) levels. However, biological diversity is concentrated in the tropics, and it is here that conservation faces the most pressing threats (Raven 1988). Furthermore, fine resolution data are often so scarce and local land-use patterns so diverse as to limit our ability to apply quantitative prioritization techniques at fine scales (Pimm and Lawton 1998). Hence, such techniques may be particularly appropriate for application in tropical areas and at continental scales. Until recently this application has been restricted to single families (Kershaw et al. 1994, 1995) or orders (Hacker et al. 1998), because continent-level species distribution data from the tropics are rarely compiled. Recognizing this limitation, the Zoological Museum of the University of Copenhagen embarked on a program to compile continent-level data on the distributions of tropical species. Such data have allowed the use of quantitative conservation prioritization techniques for birds, for which data are better than for other taxa, in both South America and Africa (Burgess et al. 1997, de Klerk 1998, Fjeldsa and Rahbek 1997, 1998, 1999). Simultaneously, the museum has compiled data from Africa for three other major taxa: mammals, snakes, and amphibians (Burgess et al. 1998). The Zoological Museum of the University of Copenhagen intends to publish these data as an atlas of African biodiversity. This article begins a series planned to extend these analyses across four major terrestrial taxa, for an entire tropical continent. For each taxon we first present an overview of patterns of species richness and narrow endemism across the continent. Second, we use these patterns to identify areas of high conservation priority that can most efficiently represent each group of species. Finally, we repeat the analysis for threatened species (Baillie and Groombridge 1996), the immediate priorities for conservation in Africa. Numerous important issues remain to be addressed, and we conclude the article with a detailed discussion of them. In our ongoing research we are address four in particular. First is the question of surrogacy—how well do conservation priorities for one taxon represent other taxa? Second is the matter of incorporating socioeconomic variables into the analyses, both as pressure (e.g., human population) and as response (e.g., existing protected areas) variables—the need to do so is urgent. The third issue involves scale: We are using environmental models to increase the resolution of the databases to scales relevant to local (rather than continental) conservation planning. The last issue concerns consensus building: The databases and analyses should become available to ongoing regional, national, and local conservation planning processes in Africa through workshops and collaborative work. The consensus built by these processes is essential for ensuring that any of these studies mentioned above results in the effective conservation of biodiversity. We compiled data on the distribution of all currently recognized 1921 species of birds, 940 species of mammals, 406 species of snakes, and 615 species of amphibians found in mainland sub-Saharan Africa (south of 20 degrees North) onto a one-degree grid, with each cell approximately 105 km on each side. We chose the one-degree resolution as a compromise between the sampling inadequacy of finer resolutions and the loss of biogeographic detail (especially in mountainous regions) at coarser resolutions (Rahbek and Graves 2000). For the 1957 one-degree grid cells across the continent holding data, we currently have a total of 828,506 species-in-grid-cell data entries for the 3882 species overall. This is a dynamic database into which new grid-cell data entries are being added almost daily; it is the most complete cross-taxonomic species distributional database for any tropical continent. Hereafter we refer to these species collectively as “African terrestrial vertebrates” for convenience, although we have not been able to include lizards in our database. For the larger and better-known species, the data are estimates of recent distributions, which we have taken from standard compilations and modified using the specialist literature where necessary. For smaller and less well-known species, we interpolated expected distributions by assuming a continuous distribution between confirmed records within relatively uniform suitable habitat, using available information on species' habitat associations and taking care to exclude known gaps in distribution. We checked interpolation by consulting taxon specialists (all sources are available at the Web site www.zmuc.dk/commonweb/research/blueprint-africa.htm). For the least well-known species, records are plotted without interpolation. Figure 1 shows examples of each of these maps. Over half of the maps are primary sources that have never been previously published or analyzed. We recognize that such binary maps represent varying degrees of simplification of actual distributions (Freitag et al. 1996), but the urgency of systematic conservation across Africa is too great to wait until better data are available (van Jaarsveld et al. 1998a). For mammals, our taxonomy follows Wilson and Reeder (1993), modified according to recent species descriptions and taxonomic opinions. For the Galagonideae we follow the species names, order, and distributional data presented by Kingdon (1997). For the larger mammal species, we produced range maps using standard references for the area (Dorst and Dandelot 1970, Haltendorf and Diller 1977, Skinner and Smithers 1990). We refined these using numerous sources, including action plans and unpublished museum data from the International Union for the Conservation of Nature and Natural Resources—Species Survival Commission. For the smaller species we compiled maps from the literature (hundreds of references) and discussed maps with dozens of specialists. For birds, our taxonomy follows Sibley and Monroe (1990, 1993), which in spite of some problems (Mayr and Bock 1994) is the most recent and comprehensive treatment of the world's avifauna. We include all terrestrial species and waterbirds that breed in the Afrotropics or that regularly visit this region as nonbreeding migrants, but exclude pelagic and vagrant species (Dowsett and Dowsett-Lamaire 1993, Dowsett and Forbes-Watson 1993). We compiled base distribution maps from the published volumes of the The Birds of Africa (Brown et al. 1982, Urban et al. 1986, 1997, Fry et al. 1988, Keith et al. 1992, Fry and Keith 2000). These base maps were then checked, refined, and completed for unmapped species using four major sources (Hall and Moreau 1962, 1970, Snow 1978, Stattersfield et al. 1998) and dozens of other published references (including recent atlases) and unpublished research. For snakes, we based our species list on Welch (1982), with the addition of newly described species and recent taxonomic reinterpretations. Data for species maps were compiled from an extensive review of the literature and visits to many museums. We used range maps taken from published compilations for poisonous snakes (Sprawls and Branch 1996) and for southern Africa (Branch 1998). For amphibians, our taxonomy follows Frost (1975) and Duellman (1983). We updated this list where necessary using recently published papers and prevailing taxonomic opinions. Data for the maps come from a thorough literature review along with unpublished data provided by numerous specialists. We used WORLDMAP (Williams 1996) to assess richness, endemism, and complementarity within our data. Stacking the species maps on top of each other provides us with richness maps (Figure 2a–d), which show the variation in numbers of species within each of the four groups across the continent. Broadly, areas with higher levels of energy-related variables such as primary productivity, potential evaoptranspiration, solar radiation, temperature, and rainfall tend to have higher species richness, as expected (Currie 1991, but see Rosenzweig and Sandlin 1997). More specifically, mammals are most speciose in the forest-dotted crescent around the north of Lake Victoria. Outliers of particular richness run south through the mountains of Kenya, Tanzania, and Malawi and west through the forest–savanna ecotone north of the Congo Basin into the Guinea forests. Areas of lowest richness lie in the Sahel, the Horn of Africa, and the Kalahari, Namib, and Karoo drylands. Birds show a richness pattern remarkably similar to that of mammals, with the main differences lying in the relatively more speciose miombo (Brachystegia, Julbernardia, and Isoberlinia) woodland of Zambia, and in the relatively poorer northern savanna–rain forest ecotone. Snakes show a different pattern. The Zambezian region contains the most species, with large numbers concentrated also along the lower Congo River and in southern Cameroon, the Albertine Rift, the coast of Kenya, and Namibia. The major drylands again have the lowest richness, but much of Ethiopia, Tanzania, and Angola are also depauperate. Amphibians are most speciose in the Lower Guinea forests, especially in Cameroon. The Albertine Rift, the Eastern Arc, and the southern African forests also hold many species, while northern Kenya and northern Tanzania, as well as the major drylands, hold rather few. The central portion of the Congolian rain forest is also quite depauperate for all four taxa. One important limitation of the maps is sampling effort. This is particularly obvious for the snakes and amphibians, as indicated by the spotty appearance of their richness maps. Often the presence of spots of particularly high species richness surrounded by apparently depauperate areas merely indicates that only a single locality has been surveyed. This effect is noticeably variable geographically; an obvious region that has been only poorly sampled is northern Mozambique (Ryan et al. 1999). However, although this problem should not be downplayed—and indeed we are taking steps to solve it by supporting both deductive environmental modeling of species ranges and the collection and compilation of new data (da et al. are that it is to the of our patterns of species richness to be by environmental within (Currie 1991, and patterns of species distributions of species with be more for the and 1997). Furthermore, range may well be of for conservation (Hall and Moreau because species with ranges tend also to have (Brown and of on both (Pimm and 2000). this Stattersfield et al. the distributions of all species with ranges of less than km to identify the of the of this is that it species with ranges than the and 1998). this an is to as a continuous for by the of the range of all species in each so that all species in the data some to the (Williams et al. Figure shows maps of In as with species richness, the most areas lie in the forests while the most are in drylands and For for with ranges are concentrated in forest around the in the Guinea forests, the Albertine Rift, the Kenya the Eastern Arc, and The and the are but other drylands, including the Horn of Africa and the are rather are concentrated in the of Africa, but with relatively species with ranges in Africa and quite along the Angola The extensive of Africa is noticeably of endemism in snakes are less well but areas holding numerous species can be in Cameroon, the Albertine Rift, the and the Horn of Africa. of southern central Africa is for snakes, but southern Africa is quite The richness and endemism maps can be using a (Williams and 1998), as in Figure regions are high in both species richness and endemism, the and have both richness and to be high endemism to species richness in the Guinea forests, the Horn of Africa, and the and In the and Zambezian have endemism, their large numbers of species. can we use these patterns of to assess which areas are the priorities for the area available for the issue can be as a question of a of priority areas one-degree grid which is of sub-Saharan we can a of areas to our of conservation this terrestrial species. This is a problem et al. a which to to this we many species in each major taxon on be in grid cells at from the maps. We of cells only of numbers of species of this range of the These are in the for areas that species be to cells that hold the numbers of species. These cells tend to be mammals, birds, and snakes, in the Albertine and for in the (Figure a tend to hold similar of species (Williams et al. and so our total within of the cells for each group is For birds, it is better than cells at be to areas that hold large of species with The most for this is to the grid cells that have the (Figure These areas are less than for species a the most cells a better of all four major than the cells we can do much better than this and by the of The of complementarity is to a of conservation areas that than a of species the of et al. 1993). We have used the first in this which is to the grid cell that the of species. In area we then all species within this first cell and as our cell the one holding the of species in the first we all species within this cell to us to where our third cell and so on until we our We the first three steps of this for mammals in Figure these steps for each major taxon us results that we can with our other methods complementarity us an to the question of where to site a conservation This is complementarity because it the area holding the of species at Numerous other methods that similar levels of but areas different from by complementarity (Csuti et al. 1997). of the problem is to for a rather than an for we to represent all of our species at least in as grid cells as a of do we repeat the steps in Figure until we have species in each We this in Figure with the at each by complementarity for the of of cells at The total of cells to species from for snakes and for to for amphibians and for birds, being the most speciose of the four taxa, can be in cells than most This can be to the that a smaller of species have restricted known ranges than do the other three (Figure in our database have a range of mammals snakes and amphibians only This effect is by three the larger range of of their our better of their distributions, and our use of continuous range maps to represent their distributions in our The cells are to represent all snakes than to represent all birds, the much of in of is that are so many species than are the to conservation a of priority it not any information about the urgency of conservation in any particular area than to many species each area the For priority areas to their of species into the the processes these species be taken into 1998). This issue has not been extensively to but one to it is to of species that are threatened with a of in the in the (Baillie and Groombridge This is by a of with distributional the conservation of most groups has been in Africa only three species of snakes and species of amphibians have been as of which all but two and from the Guinea are from South Africa (Baillie and Groombridge The conservation of mammals (Baillie and Groombridge 1996) and et al. has been In Figure we the richness of threatened African mammals and The differences from Figure richness maps for all quite Figure is more similar to Figure endemism because range is one of the for on the list (Baillie and Groombridge For mammals, the Horn of Africa, and as holding many threatened for birds, Ethiopia, the Angola and the South African hold numbers of threatened species. In the Guinea forests, the Albertine Rift, and the Eastern hold numerous threatened species. We can identify important areas for the conservation of these threatened species using the complementarity that we used for all species. represent all threatened sub-Saharan African mammal species one-degree grid the threatened can be in only We can then use these areas as a by the of cells to represent all species of mammals, and the for that we only cells to represent all species of mammals and only more for This that areas to represent all species, with the threatened species, is not less a than areas to represent species without for their conservation For mammals the cells and for the for the cells and the for incorporating be to some as habitat of the processes across Africa. The be into a of conservation priorities (Figure the top priorities as threatened areas that important to a and Such data are for through the site areas are the priorities for conservation in the top areas in to represent threatened mammals and birds, and also to represent all birds, mammals, snakes, and The represent between half and (all of the areas within the much priority areas by conservation et al. 1998, and 1998, Stattersfield et al. 1998). However, within these than of the mainland sub-Saharan biogeographic et al. are The most of these are areas in central Angola and that is by the presence in these northern areas of and and the woodland of northern the forest of central similar areas are again and Arc, Guinea forests (especially Albertine Rift, and of the areas in are in Figure also all in a for all species These are the priorities for the diversity of African terrestrial This that Arc, and Lower Albertine Rift, and as the top conservation priorities across Africa. this data be because of of the distributions of mammals and because of of for poorly known species. The of our data to terrestrial that we important areas for other taxa. third is that to mammals and is a for to biodiversity. We also do not into our area which conservation priorities within that the regions are as the priorities for the of different threatened that these are are the steps for conservation planning in We can The first is to address the question of well conservation priorities for the groups with relatively distributional data for other and indeed for as a This has been an issue of is high between groups et al. 2000). fine to be between diversity patterns et al. 1998) or conservation priorities (van Jaarsveld et al. However, from indicates that areas for most major other groups well indeed et al. 1998). The is (Burgess et al. 2000). it that are for more an may be to on environmental et al. such to habitat be with species to the loss of particular species. and is that much is in both data collection and the of both the of conservation priorities to a great (Freitag and Jaarsveld 1998). for well-known taxa, much the of comprehensive of the conservation of African snakes and amphibians, as well as the in data the of data on not particularly need is for the compilation of distribution data for because conservation priorities for in the not be well by and in the Albertine et al. 2000). This is et al. 2000). data for and conservation planning are also and should be with terrestrial data to an conservation an important research involves the use of environmental data, often from to models of expected species These techniques an to sampling gaps in species In have the potential to increase the resolution of species This research is along two of one the use of species' habitat to deductive the other the compilation of locality data to models et al. 2000). models can also information on of rather than which is important to address threats to species' and and 1998). These have been applied with great to large African mammals et al. 1999). area for is the to socioeconomic data of the resolution and as data. are some that such are at least in temperate Natural have to compile protected area data in that can be with data to to identify species that through in the et al. 1993). In Africa, analyses have been et al. but analysis has been at the continental at which species in protected areas is quite environmental have methods for into analyses and et al. 1998). numerous other socioeconomic to the of conservation have to be addressed, especially in the tropics 2000). and and and Rahbek for that areas of high endemism also hold human and of habitat this is the human and be into conservation priority examples of often and include the distributions of and and the Broadly, some methods are in to with all of these have been Finally, and most of the results of conservation need to be into analysis such as that presented here is on continental scales. The to this analysis to the national and local scales of conservation (da et al. 2000). One for so is through workshops that local with quantitative data to a consensus on conservation priorities and et al. workshops of this were in in to address the Guinea region and in in to address the Congo through such local of the planning can effective and conservation in Africa be to the de and International for to and for to and for and to de and three for with the provided by the the for the of the University of and Conservation for of maps from our a large We recent range based on published modified with numerous other data. a restricted to African drylands. data follow range maps. a We between known records within not suitable habitat and a poorly known known records are the degrees of and the indicates north of species richness and endemism in sub-Saharan African The most speciose and the forests of Lower the Albertine Rift, Kenya, the Eastern Arc, and southern in and the least speciose and most the drylands of the Sahel, the Horn of Africa, and in maps are on an for from the between species richness and (Williams and 1998). birds, snakes, and species richness is in and onto in of is then added to this to a of along which species richness and endemism areas of high species richness and endemism are in and areas of richness and endemism in In from the between richness and endemism are in with an of richness and an of richness and endemism are both high on the mountains and both in the to be high in the Guinea forests, the Horn of Africa, and the and and in the and Zambezian using for The first is to the area holding the of species, the of the species within this area are then from the a richness of species not in the The new then all the species on are The area with the richness, and the area for is all the species here indicates that the area for is Cameroon. We repeat this until the conservation is of of cells to represent all species in each major mammals, birds, snakes, and We the with through the of the of cells The total of cells to species from for snakes to for mammals distributions of range of cells for sub-Saharan African mammals, birds, snakes, and The distribution for is much less than for the other more have larger ranges of richness of threatened sub-Saharan African mammals and The most speciose Guinea forests, the Albertine Rift, the Eastern Arc, Ethiopia, and South in and the least the drylands of the and Africa and the central African rain in the grid cell areas and in as conservation priorities for sub-Saharan African The of four methods for the terrestrial species of sub-Saharan Africa in of grid by and of species. The methods are cells the the cells with the of the of the range of all species in the and complementarity cells that the of species to Conservation priorities for sub-Saharan African terrestrial species. list the first areas in for the group in We also the of species within the group by each in the The most areas lie in the Eastern the forests of and Lower the the Albertine Rift, and the follow et al.
No takes yet. Share an insight, caveat, or question.
Brooks et al. (2001) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: