boshoff.qxd The potential distributions, and estimated spatial requirements and population sizes, of the medium to large-sized mammals in the planning domain of the Greater Addo Elephant National Park project A.F. BOSHOFF, G.I.H. KERLEY, R.M. COWLING and S.L. WILSON Boshoff, A.F., G.I.H. Kerley, R.M. Cowling and S.L. Wilson. 2002. The potential distri- butions, and estimated spatial requirements and population sizes, of the medium to large-sized mammals in the planning domain of the Greater Addo Elephant National Park project. Koedoe 45(2): 85–116. Pretoria. ISSN 0075-6458. The Greater Addo Elephant National Park project (GAENP) involves the establishment of a mega biodiversity reserve in the Eastern Cape, South Africa. Conservation planning in the GAENP planning domain requires systematic information on the potential distri- butions and estimated spatial requirements, and population sizes of the medium to large- sized mammals. The potential distribution of each species is based on a combination of literature survey, a review of their ecological requirements, and consultation with con- servation scientists and managers. Spatial requirements were estimated within 21 Mam- mal Habitat Classes derived from 43 Land Classes delineated by expert-based vegeta- tion and river mapping procedures. These estimates were derived from spreadsheet models based on forage availability estimates and the metabolic requirements of the respective mammal species, and that incorporate modifications of the agriculture-based Large Stock Unit approach. The potential population size of each species was calculat- ed by multiplying its density estimate with the area of suitable habitat. Population sizes were calculated for pristine, or near pristine, habitats alone, and then for these habitats together with potentially restorable habitats for two park planning domain scenarios. These data will enable (a) the measurement of the effectiveness of the GAENP in achieving predetermined demographic, genetic and evolutionary targets for mammals that can potentially occur in selected park sizes and configurations, (b) decisions regard- ing acquisition of additional land to achieve these targets to be informed, (c) the identi- fication of species for which targets can only be met through metapopulation manage- ment, (d) park managers to be guided regarding the re-introduction of appropriate species, and (e) the application of realistic stocking rates. Where possible, the model predictions were tested by comparison with empirical data, which in general corrobo- rated the predictions. All estimates should be considered as testable hypotheses. Key words: conservation planning, mammals, distribution, density, population esti- mates, Addo, South Africa. A.F. Boshoff, G.I.H. Kerley and S.L. Wilson, Terrestrial Ecology Research Unit, Depart- ment of Zoology, University of Port Elizabeth, Port Elizabeth, 6013, Republic of South Africa; R.M. Cowling, Terrestrial Ecology Research Unit, Department of Botany, Uni- versity of Port Elizabeth, Port Elizabeth, 6013, Republic of South Africa. ISSN 0075-6458 85 Koedoe 45/2 (2002) Introduction In November 2000, the Global Environment Facility (GEF) approved a grant to South African National Parks (SANParks) to research and prepare a full proposal to the GEF for the planning and establishment of a “greater” Addo Elephant National Park (GAENP). The SW boundary of the Addo Elephant National Park (AENP) is some 35 km east of the city of Port Elizabeth (33°58'S, 25°31'E), South Africa. The vision for an expanded Addo Elephant National boshoff.qxd 2005/12/09 09:46 Page 85 Park was developed and documented by Kerley & Boshoff (1997, 2002, www.zoo. upe.ac.za/teru). Systematic conservation planning forms an integral and critical com- ponent of the implementation of the GAENP project, managed by South African National Parks (“http://www.addoelephantpark.com” www.addoelephantpark.com). Note that the original boundary of the proposed expansion to the AENP (Kerley & Boshoff 1997) has been modified, by SANParks, for the pur- poses of the conservation planning exercise for GAENP, by the addition of a 5-km buffer that follows cadastral boundaries. While the impressive plant diversity remains a major focus of conservation planning in the establishment of a GAENP (Kerley & Boshoff 1997), other biota and ecological processes which impact on the park’s biodi- versity must be taken into account in attempting to achieve its broad conservation objectives. The species’ patterns and ecolog- ical and evolutionary processes in the GAENP planning domain include the medi- um to large-sized mammals and the process- es that they drive, many of which (a) are in need of conservation intervention, and (b) may have an important impact on the park’s biota, at the species, community and ecosys- tem functioning levels. Herbivory is known to have an impact on the species composi- tion, structure and dynamics of fynbos vege- tation (Campbell 1986; Johnson 1992) and thicket vegetation (Barratt & Hall-Martin 1991; Johnson et al. 1999; Lombard et al. 2001; Moolman & Cowling 1994; Penzhorn et al. 1974; Stuart-Hill 1992; Stuart-Hill & Aucamp 1993). The important role that the proposed mega reserve will play in conserv- ing a diverse array of larger mammals, including the top predators and a number of megaherbivores, is emphasised by Kerley & Boshoff (1997). The numerous ecological processes that are mediated by the larger mammals, or that they participate in, are reported on elsewhere (Boshoff et al. 2001a). The medium to large-sized mammals were selected as “target” species (sensu Wilcox 1982) for the GAENP planning exercise because it is likely that if their minimum area requirements are met, adequate survival con- ditions will be simultaneously met for other biota. In this regard, many of these mammals qualify as “umbrella” species (sensu Wilcox 1982) since their minimum area require- ments are likely to be at least as comprehen- sive as those for the remainder of the com- munity. Mammals with a large body size (e.g., some ungulates) or which occupy a high trophic level (e.g., carnivores) are regarded as good candidates for target species acting as “umbrella” species (Wilcox 1982). In addition, the distributions and spa- tial requirements/densities of the larger mammals are probably better known, or can be better estimated, than those of the small- sized mammals in the GAENP planning domain. In any case, realistic data for these two population parameters are essential for any conservation exercise that deals with the establishment and maintenance of minimum viable populations of the larger mammalian fauna (Caughley 1994; Caughley & Sinclair 1994; Lande & Barrowclough 1987). An additional consideration for determining minimum area requirements for preserving biological diversity is that of the estimation of minimum viable populations (MVP) for “target” species (Wilcox 1982; Soulé 1987). The MVP is a set of specifications concern- ing the size and structure of the populations of a species that is necessary to provide a margin of safety from extinction. The MVP for a species can be translated into the mini- mum area requirements by determining the amount and type of habitat that will satisfy the MVP. In view of this, it is necessary for realistic estimates of the spatial require- ments/densities of each the selected species in the GAENP planning domain to be obtained. In summary, systematic data and information are required to enable conservation planners to calculate the potential numbers of individ- uals of each mammal species, within the mammal habitats within various park config- uration scenarios. These data will enable planners to measure the effectiveness of the proposed GAENP in achieving predeter- Koedoe 45/2 (2002) 86 ISSN 0075-6458 boshoff.qxd 2005/12/09 09:46 Page 86 mined demographic, genetic and evolution- ary targets for medium to large-sized mam- mals that can potentially occur in the park. In addition, they will inform decisions regard- ing acquisition of additional land, where necessary, to achieve these targets, and help identify species for which targets can only be met through metapopulation management. APPROACH The indigenous mammal species included in this study (Table 1) are those with a mass greater than ca. 2 kg (cf. Chew 1978), that are the most prominent on the landscape, and which are generally amenable to direct man- agement. As part of a separate exercise, 43 Land Classes were delineated through field mapping by Kruger & Sykes (2002), using as a basis the hierarchical classification of Subtropical Thicket by Vlok & Euston- Brown (2002). It was considered impractical to use this detailed classification for deriving the potential distributions and estimated spa- tial requirements/densities of the larger mammals. It was consequently decided to collapse the 43 Land Classes into a practical number (21) of Mammal Habitat Classes (MHCs) and to use these as the biodiversity surrogates for the mammal conservation planning component of the Greater Addo Elephant National Park. Only those Land Classes that exhibited a generally high degree of similarity, in terms of vegetation structure (and hence mammal habitat) and productivity (determining mammal densi- ties), and for which any differences that exist are considered unlikely to impact signifi- cantly on the known and potential presence and densities of mammal species, were com- bined. The potential distributions and esti- mated spatial requirements of the two otter species are based exclusively on aquatic habitats, i.e., coastline and rivers, where appropriate. DISTRIBUTIONS Methods Two steps were followed in determining the poten- tial distribution of each species, within each MHC in the GAENP planning domain. 1. Collation and interpretation of evidence that a species occurred, or could poten- tially occur in all, or in a specific part, of the GAENP planning domain. The early and recent published literature was con- sulted, as were conservation scientists and managers with a good knowledge of the macro fauna of the existing Addo Elephant National Park (AENP) and close environs (see Boshoff & Kerley 2001 and Boshoff et al. 2001b for details of the methods used). The mammal checklist for the AENP was also con- sulted, as were the mammal collection registers of the Amatole Museum in King William’s Town, where the terrestrial mammal collections from the four provincial museums in the Eastern Cape are now housed. The present study attempts to reconstruct the distrib- utions of indigenous herbivores in the period prior to arrival of European settlers, in the GAENP planning domain, in the mid 17th century. These distributions thus represent a situation where the patterns and processes exhibited by the mammals of the region were presumably still fairly intact. Thus, domestic herbivores, maintained by Khoi pastoralists in the period prior to European settlement, have not been taken into account in this analysis, owing to a lack of information on their distributions, nomadic move- ments and densities. Zoological and explorer’s records from the 17th, 18th and 19th centuries have been well reviewed by Du Plessis (1969), Rookmaaker (1989) and Skead (1987). These reviews were useful in determining the general presence or absence of most species in all or parts of the GAENP planning domain, but they generally proved to be vague in terms of the exact areas and habitats occupied by the various species. This resulted mainly from the fact that most early hunters and naturalists only recorded mammal occurrences along well travelled, or passable, routes, and few travelled at night, thereby missing the noc- turnal species. Other problems arose with interpret- ing the early, published accounts with regard to the accurate identification of some species (see Skead 1987). ISSN 0075-6458 87 Koedoe 45/2 (2002) boshoff.qxd 2005/12/09 09:46 Page 87 The following additional sources were consulted for information on the historical occurrence of mammals in the broader area around the GAENP: Coetzee (1979); Hewitt (1931); Lloyd & Millar (1983); Shortridge (1942); Skinner & Smithers (1990); Smithers (1986); Stuart (1981); Stuart (1985); Stuart et al. (1985). A review of the recent (20th century) literature revealed that surveys are incomplete in terms of species and/or area covered and tend to use political boundaries rather than ecological zones as the basic mapping units. The scale of the distribution maps in the standard account of the mammals in the southern African sub-region (Skinner & Smithers 1990) allows only generalised ranges (extents of occur- rence) to be determined. Similarly, distributions of threatened mammal species are illustrated on a broad regional basis (Smithers 1986). Museum specimens and records provide useful point data but are biased in that they only provide “presence” data, i.e., they do not represent the results of systematic data col- lection throughout the GAENP planning domain, and they do not take into account the possible migra- tory or nomadic patterns of some species. 2. Estimation of potential presence of the species, based on their ecological requirements The potential presence/absence of each species in each MHC was determined according to our under- standing of their ecological requirements, including a review of published habitat requirements (in the general GAENP area and further afield), our person- al field knowledge, and the respective habitat char- acteristics of each MHC. These characteristics included dominant plant species, vegetation struc- ture, grass component, soil nutrients, geology, topog- raphy, modal altitude, mean rainfall and rainfall sea- sonality). See Boshoff & Kerley (2001) and Boshoff et al. (2001b) for details of the methods used. As part of this exercise, AENP conservation scientists and managers, with ecological knowledge of mammals of the area, were consulted. The potential distribution of each species is present- ed according to three categories: - MHCs with the potential to sustain significant resident (i.e., present all year round and breed- ing) populations. In these MHCs the animals are generally homogeneously distributed across the landscape; - MHCs which may be used on a seasonal basis, or which may carry small populations in habitat refugia (i.e., patchy basis). In these MHCs the animals are generally not homogeneously dis- tributed, temporally and spatially, across the landscape; - MHCs where the species is unlikely to occur, except perhaps for vagrants or during rare and short incursions. In such cases the species was considered to be absent, and the MHC in ques- tion could not be relied upon to contribute to the conservation of that species. The hippopotamus potentially occurs, in suitable habitat, in and along major rivers and dams. These waterbodies must be perennial in nature and must contain pools at least 1.5 m deep. The distance trav- elled from watercourses to feeding grounds depends on forage availability and can vary widely. Hip- popotamus generally forage within about 1.5 km from waterbodies but will move freely up to eight or 10 km, and are known to move much further when forage is scarce (Skinner & Smithers 1990). Poten- tial hippopotamus habitat was marked on a digital terrain map produced by CSIR-Environmentek. For practical reasons, the overall distribution of this species is presented, rather than its distribution according to Mammal Habitat Class. Using GIS, those parts of MHCs that overlap with potential hip- popotamus habitat are considered as additional (i.e., additional to the original 21 MHCs) MHCs and are treated as such for the calculation of estimated spa- tial requirements and densities (see “Spatial Require- ments”). The potential distributions (that are linear in nature) of the two otter species were marked on a digital ter- rain map that identifies the major rivers and dams. A conservative approach was adopted in determining these distributions; only waterbodies that can be con- fidently classified as being perennial were included. The approach described above, which involves a simple model based on the estimated range of each species and its association with mappable environ- mental features expressed as a series of polygons, is broadly similar to that used in other studies (e.g., Butterfield et al. 1994). Results Of the 44 indigenous, non-marine mammal species that occur, or can potentially occur in the proposed GAENP (Table 1), 41 species occur exclusively in terrestrial habitats, whereas three species, the Cape clawless and spotted-necked otters and the hippopotamus are associated with aquatic habitats. Three species are omnivores, 18 are carnivores and 23 are herbivores. Of the 44 species, 35 are Koedoe 45/2 (2002) 88 ISSN 0075-6458 boshoff.qxd 2005/12/09 09:46 Page 88 ISSN 0075-6458 89 Koedoe 45/2 (2002) Table 1 The common and scientific names, foraging guild classifications and current (2001) presence of potentially occurring medium- to large-sized omnivorous, carnivorous and herbivorous mammals in the GAENP planning domain. P = Present in 2001. Taxonomic order (except for aardvark – see text) and nomenclature (scientific and common names) follow Skinner & Smithers (1990) Common name Scientific name Foraging guild Presence in 2001 OMNIVORES Chacma baboon Papio cynocephalus P Vervet monkey Cercopithecus aethiops P Porcupine Hystrix africaeaustralis P Aardvark Orycteropus afer P CARNIVORES Aardwolf Proteles cristatus P Brown hyaena Hyaenna brunnea Spotted hyaena Crocuta crocuta Cheetah Acinonyx jubatus Leopard Panthera pardus P Lion Panthera leo Caracal Felis caracal P African wild cat Felis lybica P Small spotted cat Felis nigripes P Serval Felis serval Bat-eared fox Otocyon megalotis P Wild dog Lycaon pictus Cape fox Vulpes chama P Black-backed jackal Canis mesomelas P Cape clawless otter Aonyx capensis P Spotted-necked otter Lutra maculicollis P Honey badger Mellivora capensis P HERBIVORES African elephant Loxodonta africana Mixed feeder P Black rhinoceros Diceros bicornis Browser P Cape mountain zebra Equus zebra zebra Bulk grazer P Burchell’s zebra Equus burchelli Bulk grazer P Bushpig Potamochoerus porcus Mixed feeder P Warthog Phacochoerus aethiopicus Concentrate grazer P Hippopotamus Hippopotamus amphibius Bulk grazer P Black wildebeest Connochaetes gnou Concentrate grazer Red hartebeest Alcelaphus buselaphus Concentrate grazer P Blue duiker Philantomba monticola Browser P Common duiker Sylvicapra grimmia Browser P Springbok Antidorcas marsupialis Mixed feeder P Klipspringer Oreotragus oreotragus Browser P Oribi Ourebia ourebi Concentrate grazer Steenbok Raphicerus campestris Browser P Grysbok Raphicerus melanotis Browser P Grey rhebok Pelea capreolus Concentrate grazer P African buffalo Syncerus caffer Bulk grazer P Kudu Tragelaphus strepsiceros Browser P Bushbuck Tragelaphus scriptus Browser P Eland Taurotragus oryx Mixed feeder P Reedbuck Redunca arundinum Concentrate grazer Mountain reedbuck Redunca fulvorufula Concentrate grazer P boshoff.qxd 2005/12/09 09:46 Page 89 Koedoe 45/2 (2002) 90 ISSN 0075-6458 already present and nine could be considered for re-introduction. The 21 MHCs delineated for this study are mapped in Fig. 1 and listed in Table 2. The potential occurrence of each species in each MHC, on a “resident” or “seasonal/patchy” basis (Table 2), is illustrated in a series of distribution maps (Figs. 2–44). Due to a lack of detailed habitat information, the hip- popotamus and the two otter species are con- sidered to be potentially resident in all habi- tats mapped for these species. Discussion Notwithstanding the constraints inherent in the approach used here, the maps provided in this report are considered to represent realis- tic potential distributions of the medium to large-sized mammals in the GAENP plan- ning domain. We stress, however, that these data are underpinned by putative habitat- mammal relationships that are testable in the future. Nonetheless, the data provide new information that is essential for effective conservation planning in the GAENP, and for developing a greater understanding of the larger terrestrial vertebrates as indicators of environmental change in the proposed park (Macdonald 1992). The black wildebeest is the only species for which the greater part of its distribution range within the GAENP planning domain falls within the 5-km buffer zone. For two other species, namely oribi and reedbuck, a significant proportion of their distribution range within the planning domain falls with- in the 5-km buffer. It is emphasised that the allocation of species to specific MHCs should not be interpreted to imply that the distributions of the mam- mals are spatially and temporally fixed in the planning domain. Because of the dearth of ecological information from the region, any reconstruction of the demographics and dynamics of the medium to large mammal populations must be based on the collection of new information. Owing, in part, to the expansion of the AENP, there are currently seven extralimital species in the park, namely gemsbok Oryx gazella, impala Aepyceros melampus, water- buck Kobus ellipsiprymnus, blesbok Damaliscus dorcas phillipsi, blue wildebeest Connochaetes taurinus, red lechwe Kobus leche and nyala Tragelaphus angasii. It is recommended that these species be removed from the AENP, in view of the real and potential ecological and economic costs of keeping them in the park (Castley et al. 2001). SPATIAL REQUIREMENTS Methods The estimated spatial requirements of each species, and the associated density estimates, refer exclusive- ly to those MHCs in the GAENP planning domain where the species is likely to occur, on a “resident” or “seasonal/patchy” basis. Omnivores and carnivores The overall lack of information from the GAENP domain precluded an estimation of the spatial requirements of the omnivores and carnivores according to individual Mammal Habitat Classes. Consequently, the planning domain was treated as a homogeneous unit for this purpose. This is likely to be more appropriate for the smaller species than for the larger ones; the abundance of the latter will gen- erally reflect the abundance and spatial distribution of the larger herbivores. Estimates of the spatial requirements of each species in each MHC were based on a review of available information on densities, social structures, breeding units, territory sizes and home ranges. However, since published ecological information for the region is not available (cf. Boshoff et al. 2001b) for any of the species that can potentially occur there, estimates based on the interpretation and extrapolation of information on the relevant species from other regions in South Africa, mainly the Nama-Karoo, Grassland and Savanna biomes (sensu Low & Rebe- lo 1996), were used as surrogates. In the case of the carnivores (especially the large predators and scav- engers such as lion and spotted hyaena) the assump- tion is made that predator-prey systems are in opera- tion and that sufficient food is available. For the sake boshoff.qxd 2005/12/09 09:46 Page 90 ISSN 0075-6458 91 Koedoe 45/2 (2002) Table 2 The presence/absence of the medium- to large-sized mammals in the GAENP planning domain, according to Mammal Habitat Class (R = Resident, SP = Seasonal/Patchy) Common name Mammal Habitat Class Fo re st T hi ck et F or es t M os ai c T hi ck et S av an na M os ai c Z uu rb er g M es ic T hi ck et A dd o H ei gh ts M es ic T hi ck et Su cc ul en t T hi ck et Sp ek bo om ve ld E as te rn S pe kb oo m N oo rs ve ld W es te rn S pe kb oo m N oo rs ve ld N oo rs ve ld G ra ss y B on tv el d K ar ro id B on tv el d K ar oo T hi ck et M os ai c K ar ro id D w ar f Sh ru bl an d K ar ro id B ro ke n V el d So ur G ra ss la nd M ix ed G ra ss y Sh ru bl an d Fy nb os R ip ar ia n W oo dl an d D un ef ie ld Su nd ay s Sa ltm ar sh Chacma baboon R R R R R R SP R R R R R Vervet monkey R R R R R R R R R SP SP SP R SP SP R SP Porcupine R R R R R R R R R R R R R R R R R R R SP Aardwolf SP SP SP SP SP SP SP SP R R R R R R R SP SP Brown hyaena SP R R R R R R R R R R R R R R R R R R SP Spotted hyaena SP SP SP SP SP R R R R R R R R SP R SP R SP SP Cheetah R R R R R Leopard R R R R R R R R R R R R R R R R R R R SP SP Lion R SP SP R SP R R R R R R R R SP R SP R SP SP Caracal SP SP R R R R R R R R R R R R R SP R SP R SP SP African wild cat SP R R R R R R R R R R R R R SP R SP R SP SP Small spotted cat SP R R R R R R R R R Serval SP SP SP SP SP SP SP SP R Bat-eared fox SP SP SP R R R R R R R R SP SP Wild dog R R R R R R R R R R R R R SP R SP R SP SP Cape fox SP SP R R R R R R R R SP SP Black-backed jackal SP R R R R R R R R R R R R R R SP R SP R SP SP Honey badger SP SP R R R R R R R R R R R R R R R R R SP SP Aardvark R SP SP R R R R R R R R R R SP R R African elephant SP R R R R R R SP SP SP R R R SP SP SP R Black rhinoceros SP R R R R R R R R R R R R R R R Mountain zebra SP SP SP SP SP SP SP SP R R R Plains zebra SP SP SP R R R R R R R SP Bushpig R R SP R R R R SP SP SP R SP Warthog SP SP SP SP R R R R R SP R R R Black wildebeest SP SP Red hartebeest SP SP R R R R R SP R R SP R SP SP Blue duiker R R SP R R SP SP SP Common duiker SP SP R R R R R R R R SP SP R SP SP SP R SP R SP Springbok SP R SP R SP SP SP R R SP Klipspringer R SP SP R SP SP R Oribi SP SP SP Steenbok SP R R R SP SP R R SP P Grysbok R SP R R SP R SP SP SP R R SP Grey rhebok R R R Cape buffalo SP R SP SP SP SP SP SP SP R R R SP SP SP SP R Kudu R R R R R R R R SP R R SP R R R Bushbuck R R R R R R R SP SP SP SP SP Eland SP SP SP SP SP SP SP SP SP SP SP SP SP SP SP SP SP Reedbuck SP Mountain reedbuck R R SP R R R SP boshoff.qxd 2005/12/09 09:46 Page 91 Koedoe 45/2 (2002) 92 ISSN 0075-6458 Fi g. 1 . T he m am m al h ab ita t c la ss es in th e pl an ni ng d om ai n fo r th e G re at er A dd o N at io na l P ar k In iti at iv e. S ee te xt f or d et ai ls . boshoff.qxd 2005/12/09 09:46 Page 92 ISSN 0075-6458 93 Koedoe 45/2 (2002) of brevity, the sources of the data used to estimate the spatial requirements have not been included in this paper. Given that they occur along rivers or along the coast- line, the density estimates for the two otter species are expressed in linear terms (individuals/km). Where the distribution of the Cape clawless otter potentially overlaps with that of the spotted-necked otter, the food resources have been equally appor- tioned between them, thereby reducing the potential density of each species. A conservative approach to the estimation of the spa- tial requirements of the omnivores and carnivores in the GAENP planning domain was adopted because of the naturally, and relatively, low herbivore carry- ing capacity in some habitats, and a generally poor understanding of the ecology of the species con- cerned. This was achieved by: (a) usually adopting the lowest densities or largest territories or home ranges provided in the literature; (b) using the home range when territory size is not known; (c) basing, in appropriate cases, the estimates only on the sizes of the territories or home ranges of breeding adults—in these cases effective densities may be higher when non-territorial individuals (e.g., sub-adults, imma- tures and juveniles) are taken into account; and (d) reducing the densities in the seasonal/patchy habitats to 20 % of those calculated for the “core” habitats (Boshoff et al. 2001b). Herbivores Given the virtual absence of information on the spa- tial requirements of herbivores in the GAENP plan- ning domain, we followed a pragmatic approach in the derivation of the necessary estimates. This involves a spreadsheet model, based on forage avail- ability estimates and the metabolic requirements of the mammal species in question. The approach fol- lowed is very similar to that described by Boshoff et al. (2001b) but some adjustments have been made to accommodate the GAENP requirements and charac- teristics. Although the porcupine is predominantly a herbivore, we have treated it as an omnivore and excluded it from the spreadsheet model, since it does not fit in the conventional grazer/browser classifica- tion. The six sequential components of the model are described below: 1. Allocation of species to foraging guilds Each herbivore species was classified according to one of four foraging guilds (Table 1, adapted from Collinson & Goodman 1982), namely: bulk grazer; concentrate grazer; mixed feeder (grazer/browser); and browser. 2. Adjustment of the agricultural stocking rate The recommended agricultural stocking rates (SRs) for the respective land/agricultural units, as calculat- ed by the South African Department of Agriculture on the basis of Large Stock Units (LSUs) (Anon. 1985), were used as guidelines for estimating forage production, and ultimately the spatial requirements of herbivores within each Mammal Habitat Class. It must be emphasised that the term “spatial require- ments” normally refers to an ecological response, whereas the term “stocking rates” normally refers to an operator/manager response. The definition and use of the LSU concept to determine stocking rates for livestock and wildlife is discussed in some detail by Boshoff et al. (2001b). Where available data (cf. Stuart-Hill & Aucamp 1993) have permitted a comparison, the agricultural stocking rate broadly agrees with published empiri- cal data. Agricultural management is usually aimed at max- imising production (Morris et al. 1999), and there- fore we adopted a highly conservative approach in the calculations for the indigenous ungulates, for the purpose of sustaining populations and protecting biodiversity. This took the form of adjusting (i.e., reducing) the Department of Agriculture stocking rate applicable to each MHC by a proportion which was estimated following a subjective assessment of the biophysical attributes, as surrogates for the pro- ductivity of forage, for the MHC in question. Key surrogates here are dominant vegetation, grass com- ponent, soil nutrient status, mean annual rainfall, rainfall seasonality, modal altitude and general topography. In this way, the agricultural SRs of MHCs characterised by low productivity, low nutri- ent soils and a limited grass component, were reduced by a higher percentage than those MHCs characterised by a higher productivity, relatively higher soil nutrient status and a relatively high grass component. Thus: Adjsr = X (1+Y) (1) where Adjsr = Adjusted stocking rate, X = agricultur- al carrying capacity/stocking rate (ha/LSU), Y = adjustment value (where, e.g., 60 % = 0.4), and LSU = Large Stock Unit. boshoff.qxd 2005/12/09 09:46 Page 93 Koedoe 45/2 (2002) 94 ISSN 0075-6458 Department of Agriculture stocking rates were not available for some MHCs, nor could they be determined, owing to mapping scale differences. In these cases, stocking rates were esti- mated according to: an interpretation of the key biophysical attributes (as listed above); the stocking rates for similar MHCs; and the stocking rates for neighbouring Mammal Habitat Classes. For the purposes of the model, these adjusted stocking rates (ha/LSU) were expressed as animal unit densities (LSU/ha). 3. Allocation of animal units to foraging guilds, within MHCs The available animal units, per hectare, within each MHC (expressed as adjusted LSU/ha) were allocated to each of the four foraging guilds, where appropriate (i.e., for each guild that was represented in that MHC). To achieve this, allocations of forage (as percentages) were made for each guild within each MHC, based on subjective estimations of the proportions and nature (e.g., sweet or sour grassveld) of graze and browse, as suggested by the MHC biophysical descriptions and Figs. 2–44. The potential distribution of the different mammals in the Greater Addo Elephant National Park planning domain, according to Mam- mal Habitat Class (MHC). Solid shad- ing denotes MHCs with the potential to sustain significant resident (i.e., pre- sent all year round and breeding) pop- ulations; grey shading denotes MHCs which may be used on an ephemeral (i.e., seasonal) basis, or which may carry small populations in habitat refu- gia (patchy basis), and no shading denotes MHCs where the species is unlikely to occur, except perhaps for vagrants or during rare and short-lived incursions. Fig. 2. Fig. 3. Fig. 4. Fig. 5. Fig. 6. boshoff.qxd 2005/12/09 09:47 Page 94 ISSN 0075-6458 95 Koedoe 45/2 (2002) our personal knowledge of these habi- tats (Appendix 1). These allocations were then corroborated with the guild structures of the herbivores occurring in each Mammal Habitat Class. For example, a check was made that the distribution patterns described earlier indicated that grazers were the domi- nant herbivores in MHCs dominated by grass. For pragmatic reasons, no distinction was made between the pre- and post- Darlington Dam scenarios, i.e., the dam was considered to be a permanent feature. Since the dam itself (water area) it does not contribute any forage, it has been subtracted from the total area of Riparian Woodland that pro- vides suitable hippopotamus habitat. 4. Allocation of available animal units to individual species within foraging guilds, within MHCs For each MHC the available animal units, calculated in Step 3 above and expressed as adjusted LSU/ha, were allocated to the herbivore species within each foraging guild. Thus, where more than one species occurs within a single foraging guild within an MHC, the LSUs accorded to that guild are allocated to these species in equal proportions. This course was chosen owing to the paucity of infor- mation on resource partitioning within these guilds. 5. Adjustment for seasonality/ patchiness Species that are resident in a MHC will most likely have different forage requirements (and possibly other eco- logical requirements, e.g., availability of surface water, shelter/cover) than species that are highly spatially localised or that may only be present for a limited part of a year (i.e., nomads or migrants). Therefore, there was a requirement for the model to incorporate seasonality and habitat patchiness. This was addressed by reducing by 60 % the amount of forage Fig. 7. Fig. 8. Fig. 9. Fig. 10. Fig. 11. boshoff.qxd 2005/12/09 09:47 Page 95 Koedoe 45/2 (2002) 96 ISSN 0075-6458 allocated (expressed as adjusted LSU/ha) to seasonal/patchy species. We assumed that the amount, and indeed quality, of resources was limit- ing, rather than their seasonal avail- ability or total absence. The basis for using a value of 60 % is the same as that used for a similar study in the Cape Floristic Region, where a value of 90 % was used (Boshoff et al. 2001b). A value of 60 % was used here to reflect the probable higher and more reliable year-round forage productivi- ty. Thus, each species in each MHC is classified as “resident” or “seasonal/ patchy” (see “Distributions”). The LSUs that were “released” by a “seasonal/patchy” species were re- allocated, in equal proportions, to other species within the same foraging guild. This gives the recalculated num- ber of LSUs available to each species within a Mammal Habitat Class. In cases where other species are not pre- sent in the same guild, the “released” animal equivalents (LSU/ha) were considered as “floaters” within that MHC—to be utilised across the graze/browse spectrum by the remain- ing species in the Mammal Habitat Class. 6. Calculation of species specif- ic densities and spatial requirements, within each MHC The number of individuals of a species per ha (density), within each MHC, was calculated as follows: D = LSUrec /Sequ (2) where D = density (number of individ- uals/ ha), LSUrec = recalculated LSUs per species (as calculated in steps 1-5 above) and Sequ = species’ LSU equiv- alent. The LSU equivalents for the species follow Grossman (1991); that for African elephant follows Meissner (1982). The estimated spatial requirement for an individual of each species, within each MHC, is calculated as follows: Fig. 12. Fig. 13 Fig. 14. Fig. 15. Fig. 16. boshoff.qxd 2005/12/09 09:47 Page 96 ISSN 0075-6458 97 Koedoe 45/2 (2002) SpRqi = 1/D (3) where SpRqi = spatial requirement (ha/individual) of an individual, D = density (individuals/ha - from equa- tion 2). Constraints A limitation on the spatial require- ments of some herbivores is provided by social interaction, namely intoler- ance of conspecifics, as well as a num- ber of other constraints, e.g., presence of surface water, seasonal food avail- ability. It is known that, irrespective of the availability of forage, social and other constraints can limit the densi- ties of ungulates (e.g., see Moen 1973), and for some species the avail- ability of food, water and shelter is superseded by social factors in deter- mining densities. In this regard, the spatial requirements predicted by our model were compared, where possi- ble, with available information to investigate whether species’ social constraints had been violated. Model testing The outputs of the model were tested by comparing spatial requirement esti- mates derived from the model with published, empirically derived obser- vations of densities of species for which appropriate data are available. Such data are not available for com- plete species assemblages. Hippopotamus For those parts of MHCs where hip- popotamus can potentially occur, a separate spreadsheet model was con- structed. It differs from the model for the MHCs without hippopotamus in that hippopotamus has been inserted as an additional herbivore (it is a bulk grazer). This results in adjustments to the allocation of forage between bulk grazers within these parts of MHCs, and ultimately the densities of all species within this foraging guild. There is no published information on the densities and spatial requirements of hippopotamus in the GAENP domain, or even in the Eastern Cape, Fig. 17. Fig. 18. Fig. 19. Fig. 20. Fig. 21. boshoff.qxd 2005/12/09 09:47 Page 97 Koedoe 45/2 (2002) 98 ISSN 0075-6458 to validate the estimates provided by the model. Results Omnivores and carnivores The estimates of the spatial requirements/densities for the omnivores and carnivores are provided in Table 3. As an exam- ple of the basis for the estimation of the spatial requirements, the case of the chacma baboon is given (Box 1; cf. Table 3). The Fig. 22. Fig. 23. Fig. 24. Fig. 25. Fig. 26. Box 1: Estimation of the spatial requirements of the chacma baboon. Breeding unit/social structure Baboons are highly social, living in female bonded troops of between four and around 100- 130 individuals, with one adult male in small troops and up to 12 males in large troops; average troop size is 40 (Skinner & Smithers 1990, Apps 1996) and troop size is apparently correlat- ed with habitat quality. Breeding density/home range/ ter- ritory size Troops have home ranges but they are not territorial and rather tend to avoid other troops (Apps 1996). In the Good Hope section of the Cape Peninsula National Park home ranges of three troops of 20, 35 and 80 baboons were 9.1, 14.8 and 33.7 km², respec- tively, with home range being related to size of troop (Devore & Hall 1965). Home ranges of 400- 4000 ha have been recorded. boshoff.qxd 2005/12/09 09:47 Page 98 ISSN 0075-6458 99 Koedoe 45/2 (2002) Ta bl e 3 E st im at ed d en si tie s an d sp at ia l r eq ui re m en ts fo r se le ct ed m ed iu m - to la rg e om ni vo re s an d ca rn iv or es in ( a) “ se as on al /p at ch y” a nd ( b) “ re si de nt ” oc cu rr en ce c at e- go ri es in th e G A E N P pl an ni ng d om ai n. S ee te xt a nd B os ho ff et a l. (2 00 1a ) fo r as su m pt io ns a nd c al cu la tio ns . S ci en tif ic n am es in T ab le 1 Sp ec ie s E xt ra po la te d de ns iti es a nd s pa tia l r eq ui re m en ts E st . d en si ty ( in d. /h a) E st . s pa t./ re q. ( ha /in d. ) Se as ./p at ch y R es id en t Se as ./p at ch y R es id en t C ha cm a ba bo on 1 tr oo p of 8 0 us es 3 40 0 ha . 0. 01 29 9 0. 02 32 6 77 43 V er ve t m on ke y 25 /tr oo p, 8 tr oo ps r eq ui re d fo r 20 0 in di vi du al s at c a. 8 0 ha /tr oo p. 0. 18 51 9 0. 33 33 3 5. 4 3 Po rc up in e Fa m ily g ro up o f 5 in di vi du al s; te rr ito ry s iz e of 8 0 ha 0. 03 44 8 0. 06 25 0 29 16 A ar dw ol f M al es a nd f em al es s ha re te rr ito ri es o f up to 8 00 h a. 0. 00 13 9 0. 00 25 0 72 0 40 0 B ro w n hy ae na C la n of 4 m em be rs h as a te rr ito ry s iz e of a bo ut 2 5 00 0 ha . 0. 00 00 9 0. 00 01 6 11 25 0 62 50 Sp ot te d hy ae na A cl an o f 9 w ou ld r eq ui re a te rr ito ry o f ar ou nd 4 0 00 0 ha . 0. 00 01 3 0. 00 02 3 79 99 44 44 C he et ah E st . h om e ra ng e fo r 5 an im al s (2 m , 3 f) a t 1 00 0 00 h a, 0. 00 00 3 0. 00 00 5 33 48 0 18 60 0 w ith 7 5% o ve rl ap . 5 0 an im al s = 10 0 00 0 + 25 % f or 1 0 ite ra tio ns L eo pa rd 1 pa ir r eq ui re s a ho m e ra ng e of a bo ut 2 0 00 0 ha 0. 00 00 6 0. 00 01 0 18 00 0 10 00 0 L io n A pr id e of 1 0 an im al s (a du lts , s ub -a du lts a nd y ou ng ) 0. 00 01 2 0. 00 02 2 81 00 45 00 m ay r eq ui re a te rr ito ry o f ab ou t 4 5 00 0 ha . C ar ac al Pa ir s ha ve o ve rl ap pi ng ( by u p to 2 0% ) ho m e ra ng es o f ab ou t 6 6 00 h a. 0. 00 02 1 0. 00 03 8 47 52 26 40 A fr ic an w ild c at 1 pa ir h as a h om e ra ng e of a pp ro xi m at el y 25 0 ha . 0. 00 44 4 0. 00 80 0 22 5 12 5 Sm al l s po tte d ca t M al es a nd f em al es h av e ov er la pp in g (c a 20 % ) te rr ito ri es o f 0. 00 07 4 0. 00 13 3 13 50 75 0 ab ou t 9 00 h a. T hu s, 1 p ai r ha s a te rr ito ry o f ab ou t 1 50 0 ha . Se rv al M al es a nd f em al es o cc ur in h om e ra ng es o f up to 3 00 0 ha , 0. 00 03 7 0. 00 06 7 27 00 15 00 ta ki ng o ve rl ap in to a cc ou nt . B at -e ar ed f ox D en si ty o f 3 an im al s/ 10 0 ha 0. 01 66 7 0. 02 94 1 60 34 W ild d og 0. 01 7 an im al s pe r 10 0h a 0. 00 00 9 0. 00 01 7 10 58 8 58 82 C ap e fo x O nl y in fo rm at io n fo r hu nt in g ra ng e (u p to 5 00 h a) ; o ve rl ap pi ng 0. 00 14 8 0. 00 26 7 67 5 37 5 ho m e ra ng es . S ay 1 p r ne ed s 75 0 ha . B la ck -b ac ke d ja ck al Te rr ito ry /h om e ra ng e of a bo ut 1 10 0 ha /p r. 0. 00 10 1 0. 00 18 2 99 0 55 0 C ap e cl aw le ss o tte r 1 pe r 3. 5 km o f ri ve r; 1 p er 2 k m o f co as tli ne Sp ot te d- ne ck ed o tte r 1 pe r 2. 5 km o f ri ve r H on ey b ad ge r M al e an d fe m al e ha ve o ve rl ap pi ng h om e ra ng es o f ab ou t 1 0 00 0 ha . 0. 00 01 1 0. 00 02 0 90 00 50 00 A ar dv ar k E st . h om e ra ng es f or 1 m al e an d 1 fe m al e of 7 5 00 h a. 0. 00 01 5 0. 00 02 7 67 50 37 50 boshoff.qxd 2005/12/09 09:47 Page 99 Koedoe 45/2 (2002) 100 ISSN 0075-6458 Table 4 Estimated densities and spatial requirements of the larger mammalian herbivores in areas of 20 Mam- malian Habitat Classes in the GAENP planning domain that do not contain hippopotamus habitat. Data derived from a spreadsheet model. See text for calculations and assumptions. Scientific names in Table 1 Mammal Species Density Est. spat. Species Density Est. spat. Habitat (ind./ha) req. (ind./ha) req. Class (ha/ind.) (ha/ind.) Forest African elephant 0.00035 2876 Common duiker 0.01704 59 Bushpig 0.01758 57 Bushbuck 0.03833 26 Blue duiker 0.16611 6 Thicket Forest African elephant 0.00540 185 Common duiker 0.06617 15 Mosaic Black rhinoceros 0.00361 277 Grysbok 0.34740 3 Bushpig 0.06818 15 Cape buffalo 0.00208 482 Blue duiker 0.69481 1 Bushbuck 0.16034 6 Thicket Savanna African elephant 0.00540 185 Oribi 0.04082 25 Mosaic Black rhinoceros 0.00807 124 Steenbok 0.06122 16 Burchell’s zebra 0.00866 116 Grysbok 0.06122 16 Bushpig 0.00974 103 Cape buffalo 0.02136 47 Warthog 0.01143 88 Kudu 0.02466 41 Red hartebeest 0.00772 130 Bushbuck 0.10243 10 Blue duiker 0.12245 8 Eland 0.00198 504 Common duiker 0.14796 7 Reedbuck 0.01143 88 Springbok 0.01429 70 Zuurberg Mesic African elephant 0.00229 436 Grysbok 0.10621 9 Thicket Black rhinoceros 0.00386 259 Cape buffalo 0.00099 1011 Cape mtn. zebra 0.00168 595 Kudu 0.01180 85 Bushpig 0.02897 35 Bushbuck 0.04902 20 Blue duiker 0.21242 5 Eland 0.00182 551 Common duiker 0.07081 14 Addo Heights African elephant 0.00229 436 Common duiker 0.07081 14 Mesic Thicket Black rhinoceros 0.00386 259 Grysbok 0.10621 9 Burchell’s zebra 0.00160 623 Cape buffalo 0.00099 1011 Bushpig 0.02897 35 Kudu 0.01180 85 Warthog 0.00094 1063 Bushbuck 0.04902 20 Blue duiker 0.21242 5 Eland 0.00182 551 Succulent African elephant 0.00111 898 Common duiker 0.06092 16 Thicket Black rhinoceros 0.00332 301 Klipspringer 0.07833 13 Cape mtn. zebra 0.00076 1323 Grysbok 0.02948 34 Bushpig 0.01407 71 Cape buffalo 0.00045 2247 Warthog 0.00571 175 Kudu 0.01015 98 Red hartebeest 0.00386 259 Bushbuck 0.04218 24 Blue duiker 0.05896 17 Eland 0.00088 1134 Spekboomveld African elephant 0.00142 706 Common duiker 0.06128 16 Black rhinoceros 0.00334 299 Grysbok 0.09192 11 Burchell’s zebra 0.00069 1452 Cape buffalo 0.00042 2354 Bushpig 0.01791 56 Kudu 0.01021 98 Warthog 0.00727 138 Bushbuck 0.04242 24 Blue duiker 0.06566 15 Eland 0.00112 891 Eastern African elephant 0.00027 3753 Springbok 0.02716 37 Spekboom Black rhinoceros 0.00387 258 Klipspringer 0.03175 32 Noorsveld Cape mtn. zebra 0.00353 284 Steenbok 0.10648 9 boshoff.qxd 2005/12/09 09:47 Page 100 ISSN 0075-6458 101 Koedoe 45/2 (2002) Burchell’s zebra 0.01852 54 Cape buffalo 0.00208 482 Warthog 0.00667 150 Kudu 0.01183 85 Black wildebeest 0.00121 828 Eland 0.00069 1458 Red hartebeest 0.00450 222 Mtn. reedbuck 0.01282 78 Common duiker 0.07099 14 Western African elephant 0.0002 24587 Springbok 0.00404 248 Spekboom Black rhinoceros 0.00348 287 Klipspringer 0.02857 35 Noorsveld Cape mtn. zebra 0.00241 416 Steenbok 0.09583 10 Burchell’s zebra 0.01263 79 Cape buffalo 0.00142 706 Warthog 0.00606 165 Kudu 0.01065 94 Red hartebeest 0.00410 244 Eland 0.00056 1782 Common duiker 0.06389 16 Mtn. reedbuck 0.01166 86 Noorsveld African elephant 0.00020 5004 Common duiker 0.06366 16 Black rhinoceros 0.00347 288 Springbok 0.02037 49 Cape mtn. zebra 0.00220 454 Steenbok 0.09549 10 Burchell’s zebra 0.01157 86 Cape buffalo 0.00130 770 Warthog 0.00833 120 Kudu 0.01061 94 Red hartebeest 0.00563 178 Eland 0.00051 1944 Grassy Bontveld African elephant 0.00252 397 Springbok 0.00667 150 Black rhinoceros 0.00801 125 Oribi 0.01905 53 Burchell’s zebra 0.01221 82 Grysbok 0.02593 39 Bushpig 0.00455 220 Cape buffalo 0.00753 133 Warthog 0.01733 58 Kudu 0.00288 347 Red hartebeest 0.01171 85 Bushbuck 0.01197 84 Common duiker 0.01728 58 Eland 0.00093 1080 Karroid Bontveld African elephant 0.00210 477 Springbok 0.00556 180 Black rhinoceros 0.00617 162 Steenbok 0.03086 32 Burchell’s zebra 0.00842 119 Grysbok 0.03086 32 Bushpig 0.00379 264 Cape buffalo 0.00519 193 Warthog 0.01667 60 Kudu 0.01886 53 Red hartebeest 0.01126 89 Bushbuck 0.01425 70 Common duiker 0.02058 49 Eland 0.00077 1296 Karoo Thicket African elephant 0.00157 635 Klipspringer 0.07093 14 Mosaic Black rhinoceros 0.00301 332 Steenbok 0.02546 39 Cape mtn. zebra 0.00198 504 Cape buffalo 0.00467 214 Bushpig 0.00284 352 Kudu 0.00919 109 Warthog 0.00333 300 Bushbuck 0.01175 85 Red hartebeest 0.00225 444 Eland 0.00058 1728 Common duiker 0.05517 18 Mtn. reedbuck 0.00641 156 Springbok 0.00417 240 Karroid Dwarf African elephant 0.00030 3336 Springbok 0.03055 33 Shrubland Black rhinoceros 0.00480 208 Klipspringer 0.02381 42 Cape mtn. zebra 0.00176 567 Steenbok 0.13194 8 Burchell’s zebra 0.00926 108 Cape buffalo 0.00104 963 Warthog 0.01250 80 Kudu 0.00309 324 Red hartebeest 0.00845 118 Eland 0.00077 1296 Common duiker 0.01852 54 Table 4 (continued) Mammal Species Density Est. spat. Species Density Est. spat. Habitat (ind./ha) req. (ind./ha) req. Class (ha/ind.) (ha/ind.) boshoff.qxd 2005/12/09 09:47 Page 101 Koedoe 45/2 (2002) 102 ISSN 0075-6458 Table 4 (continued) Mammal Species Density Est. spat. Species Density Est. spat. Habitat (ind./ha) req. (ind./ha) req. Class (ha/ind.) (ha/ind.) Karroid Broken African elephant 0.00034 2919 Springbok 0.03492 29 Veld Black rhinoceros 0.00364 275 Klipspringer 0.02449 41 Cape mtn. zebra 0.00252 397 Steenbok 0.10000 10 Burchell’s zebra 0.01323 76 Cape buffalo 0.00148 674 Warthog 0.00857 117 Kudu 0.01111 90 Black wildebeest 0.00155 644 Eland 0.00088 1134 Red hartebeest 0.00579 173 Mtn. reedbuck 0.01648 61 Common duiker 0.01905 53 Sour Grassland Cape mtn. zebra 0.01905 53 Grey rhebok 0.06000 17 Red hartebeest 0.00405 247 Cape buffalo 0.00280 357 Common duiker 0.02222 45 Eland 0.00370 270 Oribi 0.02143 47 Mtn. reedbuck 0.04615 22 Grysbok 0.03333 30 Mixed Grassy African elephant 0.00085 1182 Grysbok 0.07320 14 Shrubland Black rhinoceros 0.00266 376 Grey rhebok 0.03922 26 Cape mtn. zebra 0.01494 67 Cape buffalo 0.00220 455 Red hartebeest 0.01060 94 Kudu 0.00813 123 Common duiker 0.04880 20 Eland 0.00218 459 Klipspringer 0.06274 16 Mtn. reedbuck 0.03017 33 Steenbok 0.02614 38 Fynbos Cape mtn. zebra 0.00756 132 Grey rhebok 0.03492 29 Red hartebeest 0.00172 583 Eland 0.00353 284 Common duiker 0.06349 16 Mtn. reedbuck 0.00488 205 Grysbok 0.38094 3 Riparian African elephant 0.00308 324 Common duiker 0.16931 6 Woodland Black rhinoceros 0.00924 108 Springbok 0.01429 70 Burchell’s zebra 0.01299 77 Steenbok 0.06349 16 Bushpig 0.03896 26 Cape buffalo 0.03204 31 Warthog 0.06857 15 Kudu 0.02822 35 Red hartebeest 0.01158 86 Bushbuck 0.02930 34 Blue duiker 0.12698 8 Eland 0.00198 504 Dunefield Bushpig 0.00135 743 Grysbok 0.00658 152 Common duiker 0.00439 228 Bushbuck 0.00304 329 estimated requirement of 3400 ha for a troop of 80 individuals is derived from this infor- mation. Herbivores The model’s estimations of the spatial requirements (and densities) for all species, except hippopotamus, in a single MHC are listed in Table 4. The estimations for MHCs that can potentially carry hippopotamus are listed in Table 5. The inclusion of hippopota- mus within certain MHCs has, predictably, had the effect of lowering the densities and increasing the spatial requirements of all the bulk grazers in these habitats. Thicket forest mosaic and thicket savanna mosaic can potentially carry the highest den- sities of elephant (0.54 individuals/km²). Riparian woodland and thicket savanna mosaic can potentially carry the highest den- sities (0.8–0.9 individuals/km²) of black rhinoceros, as they can Cape buffalo boshoff.qxd 2005/12/09 09:47 Page 102 (2.1–3.2 individuals/km²) and common duik- er (14.7–16.9 individuals/km²). In the case of the African elephant, a mega- herbivore, social constraints (e.g., inter-bull aggression–Kerley & Boshoff 1997; White- house & Kerley 2002) have not been violat- ed by the model’s predictions. Interaction between individuals of another megaherbi- vore, the black rhinoceros, provides a major constraint (Adcock 1994) and a general min- imum spatial (social) requirement of 200 ha/ animal has been suggested (see Hall-Martin & Knight 1994). Only four of the 16 density estimates provided by the model are above this suggested maximum density for this species; given the nature of the habitats in question, these estimates require field test- ing. It is noteworthy that in xeric succulent thicket in the Andries Vosloo Kudu Reserve, a density of as high as 1 male rhino/50 ha has been recorded (Adcock 1994). It is difficult to obtain empirical data to test the estimates provided by the spreadsheet models. There is virtually no information for the study area, and where information is available, it is normally unsuitable due to a number of constraints. For example, popula- tions of the larger mammal species in the current AENP (including the Zuurberg por- tion), and indeed in other national parks and nature reserves in the region, are not natural and are influenced by factors such as pres- ence of fencing (affecting population num- bers, and limiting natural movements), and the absence of the large predators (influenc- ing population structure and causing behav- ioural aberrations). Thus, densities may be relatively high in protected areas due to a combination of pristine, or near-pristine, habitats and absence of the larger predators. In addition, the spatial requirements of the herbivores are known to vary between habi- tats, owing to spatial variation in forage quality and availability, and shelter. Thus, until empirical studies have been conducted in the MHCs in the GAENP planning domain, testing of the model’s predictions is always going to be problematic. Notwithstanding the contraints mentioned above, an attempt was made to compare empirically obtained spatial requirement data with the predictions from the model (Table 6). With few exceptions, the data derived from the model were broadly cor- roborated for those herbivore species for which published information is available, thereby indicating that realistic values were generated by the model. It is again empha- sised that the predictions from the model should be regarded as hypotheses and should be tested through field studies and modified where necessary (see also General Discus- sion). Discussion The spatial requirement and density data generated by the model described here are considered to be realistic. They can therefore be meaningfully used in the conservation planning exercise for the larger mammals in the GAENP planning domain. These data also provide useful information for guiding conservation management decisions, for example, determining multi-species assem- blages and preliminary stocking rates of her- bivores in the proposed GAENP. It is emphasised that the estimated densities or spatial requirements refer to a situation where the entire suite of potentially occur- ring species is available and present, and the habitats in which they can occur are in an “intact” or “potentially restorable” state. Any deviation from this scenario, e.g., due to the unavailability of a species (for various rea- sons) or total habitat transformation, will require manipulation of the data and rerun- ning of the model. We recognise, however, that the model great- ly oversimplifies the highly complex intra- specific and inter-specific mammal interac- tions, and the equally complex animal-plant relationships, the latter often being influ- enced by seasonality. There are, however, no alternatives when working at this scale, and with so little ecological information avail- able for the species concerned. ISSN 0075-6458 103 Koedoe 45/2 (2002) boshoff.qxd 2005/12/09 09:47 Page 103 Koedoe 45/2 (2002) 104 ISSN 0075-6458 Table 5 Estimated densities and spatial requirements values of the larger mammalian herbivores in areas of nine Mammalian Habitat Classes in the GAENP planning domain that contain hippopotamus habitat. Data derived from a spreadsheet model. See text for calculations and assumptions. Scientific names in Table 1 Mammal Species Density Est. spat. Species Density Est. spat. Habitat (ind./ha) req. (ind./ha) req. Class (ha/ind.) (ha/ind.) Thicket Forest African elephant 0.00540 185 Common duiker 0.06617 15 Mosaic Black rhinoceros 0.00361 277 Grysbok 0.34740 3 Bushpig 0.06818 15 Cape buffalo 0.00104 963 Hippopotamus 0.00198 504 Bushbuck 0.16034 6 Blue duiker 0.69481 1 Thicket African elephant 0.005395 185 Springbok 0.014286 70 Savanna Black rhinoceros 0.00807 124 Oribi 0.040816 25 Mosaic Burchell’s zebra 0.005772 173 Steenbok 0.061224 16 Bushpig 0.00974 103 Grysbok 0.061224 16 Warthog 0.011429 88 Cape buffalo 0.011571 86 Hippopotamus 0.005527 181 Kudu 0.02466 41 Red hartebeest 0.007722 130 Bushbuck 0.102432 10 Blue duiker 0.122449 8 Eland 0.001984 504 Common duiker 0.147958 7 Reedbuck 0.011429 88 Zuurberg Mesic African elephant 0.00229 436 Common duiker 0.07081 14 Thicket Black rhinoceros 0.00386 259 Grysbok 0.10621 9 Cape mtn. zebra 0.00112 893 Cape buffalo 0.00066 1516 Bushpig 0.02897 35 Kudu 0.01180 85 Hippopotamus 0.00173 577 Bushbuck 0.04902 20 Blue duiker 0.21242 5 Eland 0.00182 551 Succulent African elephant 0.00111 898 Common duiker 0.06092 16 Thicket Black rhinoceros 0.00332 301 Klipspringer 0.07833 13 Cape mtn. zebra 0.00050 1985 Grysbok 0.02948 34 Bushpig 0.01407 71 Cape buffalo 0.00030 3371 Warthog 0.00571 175 Kudu 0.01015 98 Hippopotamus 0.00078 1283 Bushbuck 0.04218 24 Red hartebeest 0.00386 259 Eland 0.00088 1134 Blue duiker 0.05896 17 Spekboomveld African elephant 0.00142 706 Common duiker 0.06128 16 Black rhinoceros 0.00334 299 Grysbok 0.09192 11 Burchell’s zebra 0.00046 2178 Cape buffalo 0.00028 3531 Bushpig 0.01791 56 Kudu 0.01021 98 Warthog 0.00727 138 Bushbuck 0.04242 24 Hippopotamus 0.00074 1344 Eland 0.00112 891 Blue duiker 0.06566 15 Western African elephant 0.00022 4587 Springbok 0.00404 248 Spekboom Black rhinoceros 0.00348 287 Klipspringer 0.02857 35 Noorsveld Cape mtn. zebra 0.00180 554 Steenbok 0.09583 10 Burchell’s zebra 0.00689 145 Cape buffalo 0.00106 942 Warthog 0.00606 165 Kudu 0.01065 94 Hippopotamus 0.00203 493 Eland 0.00056 1782 Red hartebeest 0.00410 244 Mtn. reedbuck 0.01166 86 Common duiker 0.06389 16 Noorsveld African elephant 0.00020 5004 Common duiker 0.06366 16 Black rhinoceros 0.00347 288 Springbok 0.02037 49 boshoff.qxd 2005/12/09 09:47 Page 104 Table 5 (continued) Mammal Species Density Est. spat. Species Density Est. spat. Habitat (ind./ha) req. (ind./ha) req. Class (ha/ind.) (ha/ind.) ISSN 0075-6458 105 Koedoe 45/2 (2002) Cape mtn. zebra 0.00165 605 Steenbok 0.09549 10 Burchell’s zebra 0.00631 158 Cape buffalo 0.00097 1027 Warthog 0.00833 120 Kudu 0.01061 94 Hippopotamus 0.00186 538 Eland 0.00051 1944 Red hartebeest 0.00563 178 Karroid African elephant 0.00210 477 Springbok 0.00556 180 Bontveld Black rhinoceros 0.00617 162 Steenbok 0.03086 32 Burchell’s zebra 0.00561 178 Grysbok 0.03086 32 Bushpig 0.00379 264 Cape buffalo 0.00346 289 Warthog 0.01667 60 Kudu 0.01886 53 Hippopotamus 0.00165 605 Bushbuck 0.01425 70 Red hartebeest 0.01126 89 Eland 0.00077 1296 Common duiker 0.02058 49 Riparian African elephant 0.00308 324 Common duiker 0.16931 6 Woodland Black rhinoceros 0.00924 108 Springbok 0.01429 70 Burchell’s zebra 0.00866 116 Steenbok 0.06349 16 Bushpig 0.03896 26 Cape buffalo 0.01736 58 Warthog 0.06857 15 Kudu 0.02822 35 Hippopotamus 0.00829 121 Bushbuck 0.02930 34 Red hartebeest 0.01158 86 Eland 0.00198 504 Blue duiker 0.12698 8 The advantages and disadvantages of using the LSU approach to estimate stocking rates, is discussed in some detail by Boshoff et al. (2001b) and Boshoff et al. (2002). We con- tend that the LSU-based approach is appro- priate for estimating densities of medium to large-sized wild mammals at a mega-reserve (e.g., GAENP) scale, and that realistic val- ues, that can be used for systematic conser- vation planning in the GAENP planning domain, have been generated. An alternative to the LSU approach for calculating stocking rates is the use of a standing crop biomass of animals as an index of carrying capacity. In savanna regions, often exhibiting high rain- fall and nutrient rich soils, primary produc- tion and animal density are generally posi- tively correlated with mean annual rainfall (Coe et al. 1976). However, soil type influ- ences and further complicates this relation- ship, even in the savannas, and the biomass of large ungulates can be as much as 20 times lower on nutrient poor soils (Fritz & Duncan 1994). The fact that savannas with nutrient rich soils support different kinds of vegetation and also different types and den- sities of herbivores from those with nutrient poor soils has been emphasised by Bell (1982). Given the high regional variation in rainfall, soil type (ranging from nutrient poor to nutrient rich soils) and presumably prima- ry productivity, in the GAENP planning domain, this approach was not attempted in the present study. There is strong evidence that a high density of elephants in the “Addo bush” habitat (Spekboomveld MHC) has a negative impact on the cover, architecture and diversity of the plants (Barratt & Hall-Martin 1991; Johnson et al. 1999; Lombard et al. 2001; Moolman & Cowling 1994; Penzhorn et al. 1974; Stu- art-Hill 1992; Stuart-Hill & Aucamp 1993), summarised in Cowling & Kerley (2002). The only published recommended density for elephant in the AENP, and specifically boshoff.qxd 2005/12/09 09:47 Page 105 Koedoe 45/2 (2002) 106 ISSN 0075-6458 Ta bl e 6 C om pa ri so ns b et w ee n th e pr ed ic tio ns o f s pa tia l r eq ui re m en ts fr om th e sp re ad sh ee t m od el ( se e te xt ) an d fr om a va ila bl e em pi ri ca l d at a fo r si m ila r ha bi ta ts b ey on d th e G A E N P pl an ni ng d om ai n. S ci en tif ic n am es in T ab le 1 M am m al s pe ci es B la ck r hi no ce ro s C ap e m ou nt ai n ze br a W ar th og B lu e du ik er C om m on d ui ke r K lip sp ri ng er O ri bi E m pi ri ca l d at a( ha /in di vi du al ) A nd ri es V os lo o K ud u R es er ve : 50 h a/ in di vi du al to 2 00 h a/ in di vi du al ( A dc oc k 19 94 ). M ea n of 4 .7 i nd iv id ua ls p er b re ed in g he rd i n M Z N P (P en zh or n 19 75 ); m ea n ho m e ra ng e pe r br ee di ng h er d in M Z N P is 9 10 h a (P en zh or n 19 82 ). = 19 4 ha /in di vi du al A nd ri es V os lo o K ud u R es er ve ( So m er s 19 92 ) 12 .3 2 an i- m al s/ km ² ( fr om 8 00 a ni m al s pe r 64 93 h a) = 8 ha /in di vi du al 0. 5– 1 ha /in di vi du al ( A pp s 19 96 ) 5. 5– 8 ha /in di vi du al ( H an ek om & W ils on 1 99 1) 1. 8– 11 h a/ in di vi du al ( V on G ad ow 1 97 8) 17 h a/ in di vi du al , an d as l ow a s 20 –5 0 ha /in di vi du al (A lle n- R ow la nd so n 19 86 ). 20 –5 0 ha /in di vi du al ( R ow e- R ow e 19 91 ). E as te rn C ap e th ic ke t: es t. le ss th an 3 ha /in di vi da l (F ur st en bu rg & K le yn ha ns 1 99 6) . 11 –1 5 ha /in di vi du al ( N or to n 19 80 ) E as te rn C ap e: 1 2. 5 ha /in di vi du al ( in cr ea si ng p op ul at io n so c ou ld b e hi gh er d en si ty ) (V an T ey lin ge n & K er le y 19 95 ). 8 to 3 0 ha /in di vi du al ( R ow e- R ow e 19 88 ) an d 6. 3 to 5 7. 6 ha /in di vi du al ( E ve re tt et a l. 19 91 ). Pr ed ic tio ns f ro m th e m od el ( ha /in di vi du al ) R an ge : 1 08 -3 76 R an ge : 5 3- 13 73 R an ge : 1 5- 10 63 R an ge : 1 -1 7 R an ge : 6 -2 28 R an ge : 1 3- 42 R an ge : 3 3- 63 N ot es T he e st im at es f ro m t he m od el o ve rl ap w ith e m pi ri ca l da ta . T he e m pi ri ca l v al ue s fa ll w ith in th e m od el 's ra ng e. A V K R d en si ty c on si de re d to b e ar tif ic ia lly h ig h du e to ab se nc e of th e la rg e pr ed at or s, e sp ec ia lly li on . T he e m pi ri ca l va lu es c or re sp on d w el l w ith t he m od el 's ra ng e. T he e m pi ri ca l va lu es f al l w ith in t he m od el 's ra ng e. T he em pi ri ca l v al ue s fa ll w ith in th e m od el 's ra ng e. T he e m pi r- ic al v al ue is s im pl y an e st im at e. T he e m pi ri ca l va lu es b ro ad ly c or re sp on d w ith t he l ow er en d (i .e . h ig he st d en si ty ) of th e m od el 's ra ng e. T hi s em pi ri ca l s tu dy w as c on du ct ed in p ri m e ha bi ta t a nd on e w ith a re la tiv el y hi gh ra in fa ll. T he em pi ri ca lly ob ta in ed v al ue s ov er la p or c or re sp on d w ith t he m od el 's pr ed ic tio ns . boshoff.qxd 2005/12/09 09:47 Page 106 ISSN 0075-6458 107 Koedoe 45/2 (2002) Ta bl e 6 (c on tin ue d) G ry sb ok G re y rh eb ok K ud u B us hb uc k M ou nt ai n re ed bu ck R an ge : 3 -1 52 R an ge : 1 7- 29 R an ge : 3 3- 32 4 R an ge : 6 -3 29 R an ge : 2 2- 20 5 1. 3– 9. 4 ha /in di vi du al ( M an so n 19 74 ). 15 –1 52 h a/ in di vi du al ( Fe rr ei ra 1 98 4) . 15 h a/ in di vi du al ( B eu ke s 19 87 ). A dd o E le ph an t N at io na l Pa rk : m ea n of 5 27 a ni m al s in 13 6 42 h a = 26 h a/ in di vi da l ( fr om S A N P) . M Z N P: c a 10 0 an im al s in 6 50 0 ha = 6 5 ha /in di vi du al (f ro m S A N P) . A nd ri es V os lo o K ud u R es er ve (A lle n R ow la nd so n 19 80 ). = 17 k ud u/ km ² o r 5. 9 ha /in di vi du al . 20 h a/ in di vi du al ( A lle n- R ow la nd so n 19 86 ) S C ap e: 1 4– 20 h a/ in di vi du al ( Se yd ac k 19 84 ). 33 h a/ in di vi du al ( O de nd aa l & B ig al ke 1 97 9) E C ap e: 7 7 ha /in di vi du al ( St ua rt -H ill & D an ck w er ts 19 88 ). M Z N P: 4 09 a ni m al s in 6 50 0 ha = 16 h a/ in di vi du al ( fr om S A N P) . T he e m pi ri ca lly o bt ai ne d va lu es o ve rl ap w ith th e m od el 's pr ed ic tio ns . T he e m pi ri ca lly o bt ai ne d va lu es o ve rl ap o r co rr es po nd cl os el y w ith th e m od el 's pr ed ic tio ns . T he m od el p re di ct s re la tiv el y lo w d en si tie s; i t m ay b e ov er es tim at in g th e sp at ia l re qu ir em en ts , fo r re as on s no t un de rs to od b ut p os si bl y lin ke d to k ud u fe ed in g ec ol og y, or th e ku du d en si tie s in th es e re se rv es m ay b e ar tif ic ia lly hi gh d ue to a n ab se nc e of la rg e pr ed at or s. In a ll ca se s qu ot ed , th e em pi ri ca l va lu es f al l w ith in t he m od el 's ra ng e. T he M Z N P co ns er ve s pr im e ha bi ta t f or th is s pe ci es . boshoff.qxd 2005/12/09 09:47 Page 107 Koedoe 45/2 (2002) 108 ISSN 0075-6458 for the Spekboomveld habitat class, is derived from research by National Parks scientists dur- ing the early 1970s, when a den- sity of 0.4 elephants per km² was proposed (Penzhorn et al. 1974). A significantly lower density of 0.142 elephants per km² was estimated for this habitat by our spreadsheet model. The highest densities predicted by the model, namely 0.54 elephants/km², are for the Thicket Forest Mosaic and the Thicket Savanna Mosaic. The basis for the fourfold higher density of 2 elephants/km² rec- ommended by Hall-Martin & Barratt (1991), and adopted by Knight et al. (2002), is question- able (Cowling & Kerley 2002). It needs to be emphasised that even though the model has attempted to address the issue of seasonality for certain species (by reducing the amount of allo- cated forage), it is important that the GAENP be managed as a single spatial unit, in order to provide maximum opportunity for movements by nomadic or migratory species, on a year- round basis. This will cater for ecological factors such as pres- ence of surface water, seasonal food availability and the possi- ble negative effects of selective foraging on threatened plants. POPULATION SIZES Methods The potential population sizes of the 44 mammal species were calculated by simply multiplying the density esti- mate for each species with the area (in hectares) of the terrestrial habitats (i.e., MHCs), or length (km), in the case of rivers or coastline. Fig. 27. Fig. 28. Fig. 29. Fig. 30. Fig. 31. boshoff.qxd 2005/12/09 09:47 Page 108 ISSN 0075-6458 109 Koedoe 45/2 (2002) These data were calculated for two park planning scenarios, namely: slightly modified Kerley & Boshoff (1997) planning domain; and the GAENP planning domain; and accord- ing to two habitat transformation cate- gories, namely: “Intact” and “Restor- able”. Fig. 32. Fig. 33. Fig. 34. Fig. 35. Fig. 36. Box 2: Estimated population sizes of the two otter species Based on the information given in Table 3, it was decided to use a linear density of 1 individual/3 km of river or coastal habitat for both species of otter, namely Cape clawless otter and spot- ted-necked otter. Note that the trans- formation of landscapes did not affect the availability of rivers for otter con- servation, i.e., all rivers were consid- ered as “intact” habitat for otters. See “Spacial Requirements” for an expla- nation of methodology followed. Cape clawless otter 183 km of potentially suitable river and coastal habitat for Cape clawless otter alone = 61 individuals 342 km of potentially suitable river habitat equally shared with spotted- necked otter; results in 171 km avail- able for Cape clawless otter = 57 indi- viduals. Grand total = 118 individuals Spotted-necked otter 342 km of potentially suitable river habitat equally shared with Cape claw- less otter; results in 171 km available for spotted-necked otter = 57 individu- als. Grand total = 57 individuals boshoff.qxd 2005/12/09 09:47 Page 109 Koedoe 45/2 (2002) 110 ISSN 0075-6458 Results The potential population sizes, according to two park planning domain scenarios are listed for all species in Table 7. The popu- lation estimates for the two otter species are presented in Box 2. Discussion The estimates of populations sizes provided in Table 7 and in Box 2, can be used to measure the degree to which mammal population targets, that are set as part of a separate conservation planning exercise (e.g., Kerley et al. in prep.), are met by the vari- ous planning domain scenarios, taking into account different transformation categories. The estimated population sizes for two species require com- ment. First, the estimates for the klipspringer require closer scrutiny, as the values appear to be somewhat high. In the MHCs where this species has been marked as being “resident” it may be more appropriate, in a future study, to map individual habitat patches. Second, the esti- mates for the kudu may be too low. The reasons for this are not known but may be linked with this species’ particular feeding ecology. A similar pattern was observed in the analysis for the Cape Floristic Region (Boshoff et al. 2002). The estimates for some of the other smaller ungu- lates may at face value also appear to be high (e.g., grysbok: JG Castley pers. comm.). We again emphasise that the model’s outputs represent potential den- sities in intact habitats, and Fig. 37. Fig. 38. Fig. 39. Fig. 40. Fig. 41. boshoff.qxd 2005/12/09 09:47 Page 110 ISSN 0075-6458 111 Koedoe 45/2 (2002) Table 7 The estimated potential total population sizes of the medium- to large-sized mammals, according to two GAENP planning domain scenarios, and according to two transformation categories. Scientific names in Table 1 Estimated number of animals Species Modified Kerley & Boshoff (1997) GAENP planning domain planning domain Intact Restorable Total Intact Restorable Total Chacma baboon 3885 1504 5388 5477 2233 7710 Vervet monkey 60440 28480 88920 90504 46297 136801 Porcupine 15915 6328 22242 22957 10250 33207 Aardwolf 356 134 490 536 227 763 Brown hyaena 37 15 52 55 24 79 Spotted hyaena 37 17 55 56 28 84 Cheetah 3 3 6 6 4 10 Leopard 25 10 36 37 16 53 Lion 40 19 58 60 30 90 Caracal 85 36 121 126 59 184 African wild cat 1696 720 2416 2554 1189 3743 Small spotted cat 130 83 213 205 129 334 Serval 36 18 54 49 28 78 Bat-eared fox 3280 2137 5416 5291 3472 8763 Wild dog 35 15 50 52 25 77 Cape fox 282 177 458 448 282 731 Black-backed jackal 418 176 594 619 287 907 Honey badger 48 19 67 70 31 101 Aardvark 45 23 68 70 38 108 African elephant 322 141 464 523 248 771 Black rhinoceros 766 401 1167 1250 677 1927 Cape mountain zebra 851 161 1012 1083 245 1328 Burchell’s zebra 996 805 1801 1724 1324 3048 Bushpig 3316 958 4274 4845 1557 6402 Warthog 1325 820 2145 2206 1362 3569 Hippopotamus 27 29 55 32 33 64 Black wildebeest 24 8 32 43 21 64 Red hartebeest 874 460 1334 1369 755 2124 Blue duiker 24849 7824 32673 36277 12352 48628 Common duiker 14479 6843 21322 21777 11204 32981 Springbok 1666 1022 2689 2929 1815 4744 Klipspringer 3574 1293 4867 4836 1954 6790 Oribi 702 510 1213 1234 929 2162 Steenbok 7967 6213 14181 13288 9958 23246 Grysbok 21964 3693 25657 28569 6296 34865 Grey rhebok 2122 88 2210 2393 106 2499 Cape buffalo 582 414 996 1058 732 1789 Kudu 2023 1120 3143 3277 1892 5169 Bushbuck 7684 2965 10649 11577 5074 16652 Eland 340 84 424 470 144 614 Reedbuck 103 128 231 219 241 460 Mountain reedbuck 1528 423 1951 1980 639 2618 boshoff.qxd 2005/12/09 09:47 Page 111 therefore cannot be equated with current densities in transformed habitats. The data in Table 7 indicate that potential populations of herbivore species that are more prevalent in the “upland” areas are less impacted by transformed (but potentially restorable) habitats than are herbivore species that are more prevalent in “lowland” areas. For example, the steenbok is marked- ly affected by transformation in the “low- land” areas, whereas the grey rhebok is not. This pattern is understandable, given that most of the transformation has occurred in the “lowland” areas. GENERAL DISCUSSION The information generated by the present study provides realistic guidelines for the testing of population targets for medium to large-sized mammals that can potentially occur in the GAENP planning domain, and for the identification of species for which metapopulation management may be required for their conservation. It will also guide park managers regarding species that are no longer present in the general GAENP domain but that can be considered for re- introduction, and in the maintenance of real- istic densities. The information in this report Koedoe 45/2 (2002) 112 ISSN 0075-6458 Fig. 42. Fig. 43. Fig. 44. boshoff.qxd 2005/12/09 09:47 Page 112 will therefore make a significant contribu- tion to achieving the overall conservation goals of the GAENP. It is important that the estimates derived by this study be treated as hypothetical guide- lines at this stage. Thus, any management action based on these estimates should be considered experimental, should be tested through adaptive management strategies and should be closely monitored. The need to test indigenous herbivore spatial requirement/density estimates in practice, and to adapt them in the light of field experi- ence, has been mentioned elsewhere (Trol- lope 1990). In addition, the final stocking rates for these herbivores should be conserv- ative, in order to cope with unfavourable conditions (Trollope 1990). We thus advo- cate a “management by hypothesis” approach, with assumptions and predictions being explicitly tested. A major advantage of the estimates presented here is that the assumptions are explicitly quantitative and can be modified as these ideas are tested, allowing adaptive management principles and actions to be employed. The concepts of “management by hypothesis” and “adaptive management” are a generally accepted approach to dealing with management chal- lenges associated with a paucity of informa- tion (Bowman 1995; Macnab 1983; May 1991). Acknowledgements We are grateful to the following persons for provid- ing insights, information and assistance in the devel- opment of this report: Rebecca Sims-Castley (TERU, University of Port Elizabeth); Amanda Lombard (Conservation Systems, Knysna); Mike Knight, Guy Castley, Lucius Moolman, John Aden- dorff (South African National Parks); Lloyd Wingate and Fred Kigozi (Amatole Museum, King William’s Town). Guy Castley provided additional valuable comments on a draft of this paper. The GIS tasks involved in collapsing the 43 original land classes to form 21 Mammal Habitat Classes (MHC), and the calculation of the areas of the MHCs that comprised, or did not comprise, habitat for hip- popotamus, and also the areas of “Intact” and “Restorable” habitat, were conducted by CSIR- Environmentek, in a consortium with TERU, as part of a GEF co-funded contract to South African Parks to provide conservation planning services. We thank Jeanne Nel, Inge Kotze, Sarah Davies and Dirk Roux, all of CSIR-Environmentek, for their patience and co-operation in providing TERU with these GIS outputs. South African National Parks is thanked for permis- sion to publish this material. References ADCOCK, K. 1994. The relevance of “territorial” behaviour in black rhino to their population management. Pp. 82-86. In: Penzhorn, B.L. & N.P.J. Kriel (eds.). Proceedings of a Symposium on Rhinos as Game Ranch Animals. Pretoria: Onderstepoort. ALLEN-ROWLANDSON, T.S. 1980. The social and spa- tial organisation of the greater kudu (Tragela- phus strepsiceros Pallas, 1766) in the Andries Vosloo Kudu Reserve, Eastern Cape. MSc dis- sertation, Rhodes University, Grahamstown. ALLEN-ROWLANDSON, T.S. 1986. An autecological study of bushbuck and common duiker in rela- tion to forest management. Ph.D. thesis, Univer- stity of Natal, Pietermaritzburg. ANONYMOUS, 1985. The National Grazing Strategy. Pretoria: Department of Agriculture. APPS, P. 1996. Smithers’ Mammals of southern Africa: a field guide. Halfway House: Southern Book Publishers. BARRATT, D.G. & A. HALL-MARTIN. 1991. The effect of indigenous browsers on valley bushveld in the Addo Elephant National Park. Pp. 14-16. In: ZACHARIAS, P.J.K., G.C. STUART-HILL & J. MIDGLEY (eds.). Proceedings of the First Valley Bushveld Symposium. Grassland Society of South Africa. Howick. BELL, R.H.V. 1982. The effect of soil nutrient avail- ability on community structure in African ecosystems. Pp. 193-216. In: HUNTLEY, B.J. & B.H. WALKER (eds.). Ecology of tropical savan- nas. Berlin: Springer-Verlag. BEUKES, P.C. 1987. Responses of grey rhebuck and bontebok to controlled fires in coastal renoster- veld. South African Journal of Wildlife Research 17: 103-108. BOSHOFF, A.F. & G.I.H. KERLEY. 2001. Potential dis- tributions of the medium- to large-sized mam- mals of the Cape Floristic Region based on his- torical accounts and habitat requirements. African Zoology 36: 245-273. BOSHOFF, A.F., G.I.H. KERLEY, R.M. COWLING & S.L. WILSON. 2001a. Conservation planning in the Greater Addo National Park: a review of the species-, population- and spatially-driven ISSN 0075-6458 113 Koedoe 45/2 (2002) boshoff.qxd 2005/12/09 09:47 Page 113 processes involving the medium- to large-sized mammals. Terrestrial Ecology Research Unit (University of Port Elizabeth), Report No. 34. BOSHOFF, A.F., G.I.H. KERLEY & R.M. COWLING. 2001b. A pragmatic approach to estimating the distributions and spatial requirements of the medium- to large-sized mammals in the Cape floristic region, based on historical accounts and habitat requirements. Diversity and Distribu- tions 7: 29-43. BOSHOFF, A.F., G.I.H. KERLEY & R.M. COWLING. 2002. Estimated spatial requirements of the medium- to large-sized mammals, according to Broad Habitat Units, in the Cape Floristic Region, South Africa. African Journal of Range and Forage Science 19: 27-42. BOWMAN, D. 1995. Down in the forest something stirred. New Scientist (9 December) 2007: 54. BUTTERFIELD, B.R., B. CSUTI & J.M. SCOTT. 1994. Modeling vertebrate distributions for Gap Analysis. Pp. 53-68. In: MILLER, R.I. (ed.). Map- ping the Diversity of Nature. London: Chapman & Hall. CAMPBELL, B.M. 1986. Plant spinescence and her- bivory in a nutrient-poor ecosystem. Oikos 47: 168-172. CASTLEY, J.G., A.F. BOSHOFF & G.I.H. KERLEY. 2001. Compromising South Africa’s natural biodiver- sity—inappropriate herbivore introductions. South African Journal of Science 97: 344-348. CAUGHLEY, G. 1994. Directions in conservation biol- ogy. Journal of Animal Ecology 63: 215-244. CAUGHLEY, G. & A.R.E. SINCLAIR. 1994. Wildlife ecology and management. London: Blackwell Science. CHEW, R.M. 1978. The impact of small mammals on ecosystem structure and function. Pp. 167-180. In: SNYDER, D.P. (ed.) Pymatuning symposium on populations of small mammals under natural conditions. Vol. 5. Spec. Publ. Ser. Pymatuning Laboratory of Ecology. University of Pittsburgh. COE, M.J., D.M. CUMMING & J. PHILLIPSON. 1976. Biomass and production of large African herbi- vores in relation to rainfall and primary produc- tion. Oecologia 22: 341-354. COETZEE, P.W. 1979. Present distributions and status of some of the mammals of Albany. Grahams- town: Albany Divisional Council. COLLINSON, R.F.H. & P.S. GOODMAN. 1982. An assessment of range conditions and large herbi- vore carrying capacity of the Pilanesberg Game Reserve, with guidelines and recommendations for management. Inkwe 1: 1-55. COWLING, R.M. & G.I.H. KERLEY. 2002. Impacts of elephants on the flora and vegetation of subtrop- ical thicket in the Eastern Cape. Pp. 55-72. In: KERLEY, G.I.H., S.L. WILSON & A. MASSEY (eds.). Elephant conservation and management in the Eastern Cape: Workshop Proceedings. Terrestrial Ecology Research Unit (University of Port Elizabeth) Report No. 35: DEVORE, I. & K.R.L. HALL. 1965. Baboon ecology. In: DE VORE, I. (ed.) Primate behaviour: field studies of monkeys and apes. New York: Holt, Rinehart & Winston. DU PLESSIS, S.F. 1969. The past and present geo- graphical distribution of the Perissodactyla and Artiodactyla in southern Africa. M.Sc. Thesis, University of Pretoria, Pretoria. EVERETT, P.S., M.R. PERRIN & D.T. ROWE-ROWE. 1991. Responses by oribi to different range man- agement practises. South African Journal of Wildlife Research 21:114-118. FERREIRA, N.A. 1984. Various aspects of the ecology of the grey rhebuck. Pelea 3: 7-13. FRITZ, H. & P. DUNCAN. 1994. On the carrying capacity for large ungulates of African savanna ecosystems. Proceedings of the Royal Society, London 256: 77-82. FURSTENBURG, D. & M. KLEYNHANS. 1996. Kudu and bushbuck in boergoat production systems in Val- ley Bushveld: ecological interactions and con- straints. Pelea 15:5-13. GROSSMAN, D. 1991. Carrying capacity, game num- bers and optimal utilisation. Seminar: Game Ranch Management. Potchefstroom: Potchef- stroom University. HALL-MARTIN, A.J. & D.G. BARRATT. 1991. The effect of elephants Loxodonta africana on thick- et vegetation in the Addo Elephant National Park. Internal Report, South African National Parks, Pretoria. HALL-MARTIN, A.J. & M.H. KNIGHT. 1994. Conser- vation and management of black rhinoceros in South African national parks. Pp. 11-19. In: PENZHORN, B.L. & N.P.J. KRIEL (eds.). Proceed- ings of a Symposium on Rhinos as Game Ranch Animals. Pretoria: Onderstepoort. HANEKOM, N. & V. WILSON. 1991. Blue duiker Philantomba monticola densities in the Tsit- sikamma National Park, and probable factors limiting these populations. Koedoe 34: 107-120. HEWITT, J. 1931. A guide to the vertebrate fauna of the Eastern Cape Province. Grahamstown: Albany Museum. JOHNSON, S.D. 1992. Plant-animal relationships. Pp. 175-205. In: COWLING, R.M. (ed.) The ecology of fynbos: nutrients, fire and diversity. Cape Town: Oxford University Press. JOHNSON, C.F., R.M. COWLING & P.B. PHILLIPSON. 1999. The flora of the Addo Elephant National Park, South Africa: are threatened species vul- nerable to elephant damage. Biodiversity and Conservation 8: 1447-1456. KERLEY, G.I.H. & A.F. BOSHOFF. 1997. A proposal for a Greater Addo National Park: a regional and national conservation and development opportu- Koedoe 45/2 (2002) 114 ISSN 0075-6458 boshoff.qxd 2005/12/09 09:47 Page 114 nity. Terrestrial Ecology Research Unit, Univer- sity of Port Elizabeth; (internal report no. 17). KERLEY, G.I.H. & A.F. BOSHOFF. 2002. The Greater Addo National Park Initiative: linking biodiver- sity to socio-economic development. Pp. 57-66. In: PIERCE, S.M., R.M COWLING, & T. SAND- WITH. (eds.). Mainstreaming biodiversity in development: case studies from South Africa. Washington, DC: International Bank for Recon- struction and Development/The World Bank. KNIGHT, M., G. CASTLEY, L. MOOLMAN & J. ADEN- DORFF. 2002. Elephant management in Addo Elephant National Park. Pp. 32-40. In: KERLEY, G.I.H., S.L. WILSON & A. MASSEY (eds.). Ele- phant conservation and management in the Eastern Cape: Workshop Proceedings. (Terres- trial Ecology Research Unit, University of Port Elizabeth; internal report no. 35). KRUGER, L. & C. SYKES. 2002. Field mapping of land classes and land transformation. Pp. 218- 262. In: Conservation Planning Framework for the Greater Addo Elephant National Park (Part 1) – Final (internal) Report. South African National Parks. LANDE, R. & G.F. BARROWCLOUGH. 1987. Effective population size, genetic variation, and their use in population management. Pp. 87-123. In: SOULÉ, M.E. (ed.). Viable populations for con- servation. Cambridge: Cambridge University Press. LLOYD, P.H. & J.C.G. MILLAR. 1983. A questionnaire survey (1969-1974) of some of the larger mam- mals of the Cape Province. Bontebok 3: 1-49. LOMBARD, A.T., C.F. JOHNSON, R.M. COWLING & R.L. PRESSEY. 2001. Protecting plants from ele- phants: botanical reserve scenarios within the Addo Elephant National Park, South Africa. Bio- logical Conservation. 102: 191-203. LOW, A. B. & A. G. REBELO. 1996. Vegetation of South Africa, Lesotho and Swaziland. Pretoria: Department of Environmental Affairs & Tourism. MACDONALD, I.A.W. 1992. Vertebrate populations as indicators of change in southern Africa. Trans- actions of the Royal Society of South Africa 48: 87-122. MACNAB, J. 1983. Wildlife management as scientific experimentation. Wildlife Society Bulletin 11: 397-401. MANSON, J. 1974. Aspekte van die biologie en gedrag van die Kaapse grysbok Raphicerus melanotis. M.Sc. dissertation. University of Stel- lenbosch. Stellenbosch. MAY, R.M. 1991. The role of ecological theory in planning re-introduction of endangered species. Symposium of the Zoological Society of London 62: 145-163. MEISSNER, H.H. 1982. Theory and application of a method to calculate forage intake of wild south- ern African ungulates for purposes of estimating carrying capacity. South African Journal of Wildlife Research 12: 41-47. MOEN, A.N. 1973. Wildlife Ecology. San Francisco: Freeman. MOOLMAN, H.J. & R.M. COWLING. 1994. The impact of elephant and goat grazing on the endemic flora of South African succulent thicket. Biolog- ical Conservation 68: 53-61. Morris, C.D., M.B. Hardy & P.E. Bartholomew. 1999. Stocking rate. Pp. 186-193. In: TAINTON, N. (ed.). Veld management in South Africa. Pietermaritzburg: University of Natal Press. NORTON, P.M. 1980. The habitat and feeding ecology of the klipspringer Oreotragus oreotragus in two areas of the Cape Province. M.Sc. dissertation. University of Pretoria. Pretoria. ODENDAAL, P.B. & R.C. BIGALKE. 1979. Home range and groupings of bushbuck in the southern Cape. South African Journal of Wildlife Research 9:96- 101. PENZHORN, B.L. 1975. Behaviour and population ecology of the Cape mountain zebra in the Mountain Zebra National Park. D.Sc. thesis. University of Pretoria. PENZHORN, B.L. 1982. Home range sizes of Cape mountain zebras Equus zebra zebra in the Mountain Zebra National Park. Koedoe 25: 103- 108. PENZHORN, B.L., P.J. ROBBERTSE & M.C. OLIVIER. 1974. The influence of the African elephant on the vegetation of the Addo Elephant National Park. Koedoe 17: 137-158. ROOKMAAKER, L.C. 1989. The zoological explo- ration of southern Africa 1650-1790. Rotterdam: Balkema. ROWE-ROWE, D.T. 1988. Oribi - Natal’s most threat- ened antelope. Pietermaritzburg: Natal Parks Board. (Wildlife Management Technical Guides for Farmers no. 20.) ROWE-ROWE, D.T. 1991. The ungulates of Natal. Pietermaritzburg: Natal Parks Board. SEYDACK, A.H.W. 1984. Application of a photo- recording device in the census of larger rain-for- est mammals. South African Journal of Wildlife Research 14: 10-14. SHORTRIDGE, G.C. 1942. Field notes on the first and second expeditions of the Cape Museums mam- mal survey of the Cape Province and descrip- tions of some new subgenera and species. Annals of the South African Museum 36: 27-100. SKEAD, C.J. 1987. Historical mammal incidence in the Cape Province: Vol. 2 – The eastern half of the Cape Province, including the Ciskei, Transkei and East Griqualand. Cape Town: Dept Nature & Environmental Conservation. SKINNER, J.D. & R.H.N. SMITHERS. 1990. The mam- mals of the southern African subregion. Pretoria: University of Pretoria. ISSN 0075-6458 115 Koedoe 45/2 (2002) boshoff.qxd 2005/12/09 09:47 Page 115 Smithers, R.H.N. 1986. South African red data book: terrestrial mammals. Pretoria: Council for Sci- entific and Industrial Research, National Scien- tific Programmes Unit. (South African National Science Programmes report; no. 125). SOMERS, M. 1992. The implications of social struc- ture for the conservation and control of a warthog Phacochoerus aethiopicus population in the Andries Vosloo Kudu Reserve, Eastern Cape province. M.Sc. dissertation. University of Pretoria, Pretoria. SOULÉ, M. (ed.) 1987. Viable populations for con- servation. Cambridge: Cambridge University Press. STUART, C.T. 1981. Notes on the mammalian carni- vores of the Cape Province, South Africa. Bontebok 1: 1-58. STUART, C.T. 1985. The status of two endangered carnivores occurring in the Cape Province, South Africa, Felis serval and Lutra maculicol- lis. Biological Conservation 32: 375-382. STUART-HILL, G. C. 1992. Effects of elephants and goats on the Kaffrarian succulent thicket of the eastern Cape, South Africa. Journal of Applied Ecology 29: 699-710. STUART-HILL, G. C. & A.J. AUCAMP. 1993. Carrying capacity of the succulent valley bushveld of the eastern Cape. African Journal of Range and For- age Science 10: 1-10. STUART-HILL, G. & J. DANCKWERTS. 1988. Influence of domestic and wild animals on the future of succulent valley bushveld. Pelea 7: 45-46. STUART, C.T., I.A.W. MACDONALD & M.G.L. MILLS. 1985. History, current status and conservation of large mammalian predators in Cape Province, Republic of South Africa. Biological Conserva- tion 31: 7-19. TROLLOPE, W.S.W. 1990. Veld management with specific reference to game ranching in the grass- land and savanna areas of South Africa. Koedoe 33: 77-87. VAN TEYLINGEN, K. & G.I.H. KERLEY. 1995. Habitat characteristics of increasing and decreasing oribi subpopulations in the Eastern Cape Province, South Africa. South African Journal of Wildlife Research 25:118-122. VLOK, J.H. & D. EUSTON-BROWN. 2002. Subtropical Thicket Ecosystem Planning (STEP) Project: Biological Survey Report (Plants and birds). Unpublished Report, Terrestrial Ecology Research Unit, University of Port Elizabeth. VON GADOW, K. 1978. A pellet count of blue duiker and bushbuck in the Knysna forests. South African Forestry Journal 107: 77-81. WILCOX, B.A. 1982. In situ conservation of genetic resources: determinants of minimum area requirements. Pp. 639-647. In: MCNEELY, J.A. & K.R. MILLER (eds.). National parks, conserva- tion and development—the role of protected areas in sustaining society. Washington DC: Smithsonian Institution. WHITEHOUSE, A.M. & G.I.H. KERLEY. 2002. Retro- spective assessment of long-term conservation management of elephants in Addo Elephant National Park, South Africa. Oryx 36: 243-248. Koedoe 45/2 (2002) 116 ISSN 0075-6458 Fo re st T hi ck et F or es t M os ai c T hi ck et S av an na M os ai c Z uu rb er g M es ic T hi ck et A dd o H ei gh ts M es ic T hi ck et Su cc ul en t T hi ck et Sp ek bo om ve ld E as te rn S pe kb oo m N oo rs ve ld W es te rn S pe kb oo m N oo rs ve ld N oo rs ve ld G ra ss y B on tv el d K ar ro id B on tv el d K ar oo T hi ck et M os ai c K ar ro id D w ar f Sh ru bl an d K ar ro id B ro ke n V el d So ur G ra ss la nd M ix ed G ra ss y Sh ru bl an d Fy nb os R ip ar ia n W oo dl an d D un ef ie ld Su nd ay s Sa ltm ar sh Appendix 1 Adjusting stocking rates and proportional allocation of stocking opportunities to foraging guilds (see text for methods) Adjusted Stocking Rate (LSU/ha) 0.017 0.111 0.143 0.059 0.059 0.048 0.045 0.056 0.045 0.042 0.056 0.056 0.042 0.042 0.048 0.050 0.059 0.048 0.143 0.004 0.000 Bulk Grazer 1 5 20 9 9 5 5 30 25 25 29 20 15 20 25 30 20 10 30 0 0 Concentrate Grazer 1 1 20 1 1 15 10 10 10 10 18 15 15 15 15 30 20 10 15 0 0 Mixed Feeder 29 27 15 25 25 15 20 10 10 10 18 15 15 15 15 20 20 20 15 20 0 Browser 69 67 45 65 65 65 65 50 55 55 35 50 55 50 45 20 40 60 40 80 0 boshoff.qxd 2005/12/09 09:47 Page 116