Cost-benefit analysis of increasing sampling effort in a baited-camera trap survey of an African leopard (Panthera pardus) population

The use of baits at camera trap stations has been shown to increase capture rates in population surveys of large carnivores. This study set out to establish the most cost-effective density and duration of sampling for baited-camera trapping (BCT) of leopards in a semi-arid savanna environment. To determine this, we used batches of 30 BCT stations (sampling occasions) to survey a population of leopards (Panthera pardus) at Malilangwe Wildlife Reserve in south-eastern Zimbabwe from July to October 2017. We applied combinations of low to high sampling densities (2-7 occasions) and short to long sampling durations (2-14 days) and observed the effects on population estimates and cost of conducting the survey. Sixty-one leopards were identified from 4596 photographs collected over 2940 camera days. At the highest level of sampling (7 occasions and 14 days), 50 out of the 61 recorded individuals were captured more than twice indicating a near-complete survey, so the population estimate at this level (61; 95% CI = 61-67) was used as a benchmark to gauge accuracy of estimates from lower levels of effort. Accuracy and precision of population estimates stabilized over a range of efforts with 4 occasions and 9 days being the most cost-effective and sampling beyond this point incurring increased costs but negligible benefits. A minimum cost of US$47 km(-2) was required to obtain a reliable estimate with running costs responding more to changes in sampling density than sampling duration. We concluded that using BCT stations at a density of 0.24 cameras km(-2) for 9 days is optimal for censusing leopards in semi-arid savannas. Conducting a cost-benefit analysis may help researchers allocate resources and time within budget and technical constraints, ensuring that effort is not expended beyond what is economically or ecologically necessary. (C) 2019 The Authors. Published by Elsevier B.V.