Too cool, low, and slow: Camera traps used to determine leopard tortoise occupancy in arid areas of South Africa

The leopard tortoise (Stigmochelys pardalis) is the only large tortoise in southern Africa and is the most widespread of any sub-Saharan testudinid. However, the population trend of this species is unknown, and it has become increasingly rare outside fenced areas and farmland. Little is known about their behavioural ecology in response to environmental stimuli. Accurate information for quantifying activity patterns, distributions and wildlife behaviour can be obtained using camera traps. Few studies in Africa use camera traps for studying herpetofauna, especially tortoises, as they are cryptic and slow-moving, and their ectothermic nature makes using camera traps on them especially difficult. This is reflected in the use of camera trap data to study leopard tortoise occupancy from studies specifically designed to detect medium- to large-bodied mammals. Camera trap data from four arid reserves were used, where site- and observation-level covariates were used to assess leopard tortoise occupancy as a function of these covariates. Detections were sparse and uneven across reserves, producing low detectability and high uncertainty in occupancy estimates, and many top-ranked models reduced to the null. Nonetheless, the most supported models retained a small subset of covariates, notably elevation, distance to roads and outcrops, visibility, and shade, highlighting the correlates most likely to influence opportunistic leopard tortoise detections in mammal-focused camera trap surveys. This study emphasises the importance of designing monitoring programmes with specific objectives. Understanding these different factors may contribute to the knowledge concerning managing and conserving herpetofauna in arid areas.