Published Research

Hidden Markov models have become a popular time series method for the analysis of GPS tracked animals. Their advantage for identifying latent behavioural states compared with Independent Mixture models is that they take into account the time series dependency of successive displacement distances by the tracked animals. However, little is known about how the analysis results may differ depending on which of these approaches is used. We compared the results and interpretations obtained from fitting Hidden Markov and Independent Mixture models to simulated movement data as well as to field data recording the hourly movements of sable antelope and buffalo within the Kruger National Park, South Africa. Hidden Markov models consistently yielded narrower confidence intervals around parameters and smaller standard errors than simpler time independent mixture models, but for some data the improvement was marginal and the Independent Mixture model provided an adequate alternative for identifying the latent behavioural states of the animal. In general, it is expected Hidden Markov models will provide the better balance between model complexity and extensibility for animal movement modelling from a statistical perspective. However, in some cases, Independent Mixture models could provide an adequate alternative method and might be more faithful biologically.

Large carnivores are depicted to shape entire ecosystems through top-down processes. Studies describing these processes are often used to support interventionist wildlife management practices, including carnivore reintroduction or lethal control programs. Unfortunately, there is an increasing tendency to ignore, disregard or devalue fundamental principles of the scientific method when communicating the reliability of current evidence for the ecological roles that large carnivores may play, eroding public confidence in large carnivore science and scientists. Here, we discuss six interrelated issues that currently undermine the reliability of the available literature on the ecological roles of large carnivores: (1) the overall paucity of available data, (2) reliability of carnivore population sampling techniques, (3) general disregard for alternative hypotheses to top-down forcing, (4) lack of applied science studies, (5) frequent use of logical fallacies, and (6) generalisation of results from relatively pristine systems to those substantially altered by humans. We first describe how widespread these issues are, and given this, show, for example, that evidence for the roles of wolves (Canis lupus) and dingoes (Canis lupus dingo) in initiating trophic cascades is not as strong as is often claimed. Managers and policy makers should exercise caution when relying on this literature to inform wildlife management decisions. We emphasise the value of manipulative experiments and discuss the role of scientific knowledge in the decision-making process. We hope that the issues we raise here prompt deeper consideration of actual evidence, leading towards an improvement in both the rigour and communication of large carnivore science.

Rationale: Stable isotope analysis (SIA) of whiskers has been used to identify temporal feeding habits, intra‐population diet variation, as well as individual dietary specialisation of marine and terrestrial carnivores. However, the potential of the method to disclose such dietary information for large wild felids is hampered by lack of information on species‐specific whisker growth rates, whisker growth patterns and whisker‐diet trophic discrimination factors (TDFs). Methods: Whisker growth rates and growth patterns were measured for four lions (Panthera leo) and one leopard (Panthera pardus) held at the National Zoological Gardens, Pretoria, South Africa. Actively growing whiskers of the felids were 'marked' four times over 185 days using 13C‐depleted, C3‐based giraffe (Giraffa camelopardalis) meat. The periods with low δ13C values, identified following serial sectioning of the regrown whiskers at 1 mm intervals and isotopic analysis, were then correlated to specific giraffe meat feeding bouts and hence growth periods. δ13C and δ15N whisker‐diet TDFs were estimated for five lions whose diet remained consistent over multiple years.

The spotted ground thrush (Zoothera guttata; SGT) is a globally endangered forest specialist whose distribution is poorly understood in the critically endangered Indian Ocean Coastal Belt Forest (IOCBF) biome on the east coast of southern Africa. More than 70% of the IOCBF has been lost to development, resulting in fragmented forest remnants within a mosaic of different land uses. We conducted camera-trap surveys with 5796 trap-days in 82 forest patches of various sizes across a gradient of land-use types in the IOCBF during the winter seasons of 2014–2016 to establish occupancy (Ψ) of SGT. We used occupancy modelling and incorporated forest microhabitat characteristics and surrounding land-use classifications to compare sites and determine SGT distribution across the habitat mosaic. The top model estimated an average Ψ = 0.39 ± 0.09, and detection probability (P) = 0.11 ± 0.03. At the landscape scale, larger, less isolated patches had a higher Ψ of SGT while P decreased with proximity to cultivated land. At the microhabitat scale, a high percentage of bare ground influenced Ψ positively but P negatively, as did short grass cover. However, short herbaceous cover and high stem density of large (11–15 m) trees influenced P positively. These results detail the influence of landscape-scale factors including forest fragmentation and conversion to agriculture in conjunction with forest microhabitat-scale characteristics, on patch selection and distribution of a cryptic, ground-dwelling species that is both threatened and under-studied. Given its endangered status and dependence on larger and more mature forest patches, SGT may be an effective flagship species for the conservation of threatened IOCBF patches.

African catfish, Clarias gariepinus, is an important species in aquaculture and fisheries in Kenya. Mitochondrial D-loop control region was used to determine genetic variation and population structure in samples of C. gariepinus from 10 sites including five natural populations (Lakes Victoria (LVG), Kanyaboli (LKG), Turkana (LTA), Baringo (LBA) and Jipe (LJP), and five farms (Sangoro Aquaculture Center (SAN), Sagana Aquaculture Centre (SAG), University of Eldoret Fish Farm (UoE), Kibos Fish Farm (KIB), and Wakhungu Fish Farm (WKU)) in Kenya. Similarly, samples from eight localities (four natural populations: LVG/LKG, LTA, LBA, and four farmed: SAN, SAG, KIB, UoE) were genotyped using six microsatellite DNA loci. For the D-loop region, samples from natural sites exhibited higher numbers of haplotypes and haplotype diversities compared to farmed samples, and 88.2% of haplotypes were private. All except LJP and LTA shared haplotypes, and the highest number of shared haplotypes at (8) was detected in KIB. The 68 haplotypes we found in 268 individuals grouped into five phylogenetic clades: LVG/LKG, LTA, LBA, LJP and SAG. Haplotypes of farmed C. gariepinus mostly have haplotypes typical of LVG/LKG, and some shared haplotypes of the LBA population. Microsatellite analysis showed farmed samples have higher numbers of alleles than wild samples, but higher observed and expected heterozygosity levels were found in samples of natural populations. Fifteen pair-wise comparisons had significantly different FST values. All samples were in Hardy-Weinberg equilibrium. Samples from the eight localities grouped into four genetic clusters (LVG/LKG, LTA, LBA and SAG), indicating genetically distinct populations, which should be considered for aquaculture and conservation.

Adaptive governance and network governance theory provide a useful conceptual framework to guide the conservation of threatened species in complex multi-actor, multijurisdictional social ecological systems. We use principles from this theory to assess strengths and weaknesses in (1) national legislation, and (2) the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES) Regulations applicable to the conservation of the Cape mountain zebra (Equus zebra zebra) (CMZ) in South Africa.A legislated conservation tool, Biodiversity Management Plans for Species (BMP-S), establishes a collaborative network of role players and facilitates the important principles of collaborative learning and adaptation. Effective governance of this network is critical to success, but challenging because of a mandate gap and limited capacity in government to provide essential network-level competencies.National regulations governing human use ofCMZ(Threatened or Protected Species (TOPS) Regulations) accords with the principles of (1) being developed in consultation with stakeholders and (2) open to revision and adaptation. CITES Regulations also provide adequately for adaptation. Poor alignment of regulations between different regulatory authorities in South Africa and limited capacity for implementation of regulations seriously constrain learning and adaptation.

African lions (Panthera leo) are threatened across their natural range. However, the Kgalagadi Transfrontier Park (KTP) is a stronghold for the species. A population assessment in 2010 observed a skew in the sex structure with a greater proportion of males (56%) which raised concerns about the long-term sustainability of the greater KTP lion population. A key indicator of how a population responds to changes in population structure is population size. We conducted an intensive lion census between 2013 and 2015 in the southwestern KTP (14 250 km2) driving 49 784 km over 317 sampling days, which resulted in 1162 lion sightings. Registering the population through individual identification (n = 261) provided a benchmark against which other techniques could be measured and for a non-invasive marking technique. Open-population mark-recapture provided the most precise estimate of population size (n = 246; 95%CI: 237–256). Track indices (n = 242; 95%CI: 176–307) provided a similar best estimate, but were imprecise. Minimum-known-alive calculations (n = 145) provided a gross underestimate. All the methods used in this study indicate a larger lion population size than previous estimates.

The Indian Ocean Coastal Belt (IOCB) of South Africa is a natural forest-grassland mosaic, nested within an anthropogenic, mixed land-use matrix. Given the ongoing threat of agricultural expansion and urbanisation, we assessed the value of a buffer habitat (Coastal dense bush) for conserving forest species. We investigated the influence of microhabitat complexity on mammal communities within forest and dense bush habitats, using occupancy modelling. We found vertical stratification gradients as observed in studies of tropical forest chronosequence, i.e. increased foliage density in lower habitat layers and decreased foliage density in higher habitat layers for dense bush, and vice versa for forest. Structural composition suggests that dense bush is within a successional stage of secondary forest regeneration. Bushbuck (Tragelaphus scriptus) occupancy was higher in forest than dense bush, while the opposite was true for blue duiker (Philantomba monticola). Large-spotted genet (Genetta tigrina), Cape porcupine (Hystrix africaeaustralis) and marsh mongoose (Atilax paludinosus) occupancy remained constant between habitats. Grey duiker (Sylvicapra grimmia) occupancy varied greatly between dense bush (0.48 ± 0.01) and forest (0.16 ± 0.01). Dense bush appeared to maintain natural forest assemblages, and may play a crucial role in buffering IOCB forest patches, given their highly-restricted distribution. However, dense bush habitats have no protection status, but play a role in the conservation of forest plants and animals. Therefore, we advocate the inclusion of dense bush habitats in conservation networks focused on forests.

The Indian Ocean Coastal Belt Forest is extremely biodiverse but is threatened by anthropogenic land-use change. In South Africa, remnant forest patches are divided into highly restricted but protected indigenous forest (IF) and abundant but unprotected coastal thicket/dense bush (DB), which likely represents secondary/regenerating IF. We tested the hypothesis that DB had value for conserving forest-dependent avian taxonomic and functional diversity, as (a) DB is situated with in the forest-anthropogenic land-use mosaic and (b) birds maintain diverse ecosystem functions and are sensitive to environmental change given their varied biological traits. We conducted 149 and 112 fixed-radius point-count surveys of avian communities in 99, and 24, patches of DB and IF, respectively, at each location recording microhabitat vegetation structure. We compared vegetation characteristics and taxonomic species richness between vegetation classes, and tested the breadth of functional trait-space and for significant associations with vegetation structures at each survey location, using RLQ analysis. Vegetation classes differed significantly, but floral species overlap and vegetation structures in DB closely resembled regenerating IF habitat. Specialist ground-nesting and insect-gleaning species corresponded to taller, more diverse IF structures. Other biological traits and total species richness overlapped broadly across both classes, but IF had significantly higher mean species richness. Given the relatively high taxonomic and functional diversity of the avian community in the abundant DB, the scarcity of IF, and the proximity of DB to the anthropogenic land-use mosaic, we recommend allocating suitable DB patches to the Protected Area Network, which may simultaneously help preserve avian taxonomic and functional diversity, facilitate colonisation throughout the fragmented landscape and provide ecosystem services to anthropogenic land uses.

The impala (Aepyceros melampus ssp.) is a widespread antelope species occurring in sub-Saharan Africa. The two recognized subspecies have non-overlapping distribution ranges, with no known natural mixture of these subspecies until human interference. A number of common impala individuals (A. m. melampus) displaying phenotypic characteristics commonly observed in the black-faced impala (A. m. petersi), namely black facial markings, were seen on a farm in the Limpopo Province, South Africa. This farm falls outside the natural distribution range of A. m. petersi. We therefore aimed to identify the taxonomic placement of these individuals (i.e. A. m. melampus or A. m. petersi) through phylogenetic and molecular clock analyses using D-loop and cytochrome subunit b sequence data. Our results showed that these black-nosed impala from Limpopo are in fact A. m. melampus individuals. The existence of the black-nose phenotype in common impala could be more widespread than previously estimated. The occurrence of introgression between the two subspecies in this region could, however, not be fully excluded, and can only be fully assessed through the use of nuclear DNA analysis.