Estimating reliable mammal abundance in a South African commercial farming landscape using camera traps.

Abstract

With an expanding global human population, there is a rise in competition between humans and wildlife, resulting in increased negative human-wildlife interactions that impact both groups. Managing these interactions is a global conservation challenge. A lack of animal population estimates, especially outside formally protected areas, hampers effective conservation policy and management. Understanding how these species are persisting and in what number is integral to developing strategies designed to promote harmonious coexistence between humans and wildlife. This study contributes to our understanding by estimating animal densities of those often implicated in negative human-wildlife interactions in mixed South African commercial farmland, with a particular focus on camera trapping methodologies. To achieve this, three camera trap surveys over two 200 km2 areas in the Alldays area of Limpopo Province, South Africa were conducted. Distance Sampling with Camera Traps (CTDS) was applied to estimate mesocarnivore densities, resulting in estimates for African civet (Civettictis civetta) (14 and 1.7 individuals/100km2), black-backed jackal (Canis mesomelas) (46.2 and 29.8 individuals/100km2), brown hyena(Hyaena brunnea) (11.6 individuals/100km2) and caracal (Caracal caracal) (0.8 individuals/100km2); and to investigate the sensitivity of CTDS results to assumptions in determining the value of the ‘snapshot interval’. To estimate chacma baboon (Papio ursinus) densities, CTDS was used in conjunction with the Random Encounter Model (REM), achieving a density estimate of between 1-3 individuals/km2. Finally, Spatially Explicit Capture-Recapture (SECR) and occupancy modelling were employed to estimate large carnivore density and space-use. This resulted in a leopard (Panthera pardus) density estimate of 2.21 individuals/100km2 and showed that if lethal control is not ubiquitous in the landscape, chains of safer spaces will emerge that permit landscape connectivity, integral in large carnivore conservation. The results highlight the significance of mixed agricultural regions in supporting wildlife populations and fostering connections between protected areas, but also indicate potential mesopredator release in the Alldays area. Maintaining greater large carnivore abundance and connectivity might help to reduce mesopredator release, and in turn, improve human-wildlife coexistence. Some density estimates were quite imprecise and future methodological work should seek to better understand these precision issues. In the meantime, SECR should continue to be the primary method used for animals with natural unique markings, while REM will likely provide more precise density estimates of unmarked animals.