Epsilon Archive for Student Projects

Abstract

Reliable population densities are vital for effective wildlife management and conservation strategies. In this context, camera traps have become a popular tool for wildlife managers and researchers. The Random Encounter Model (REM) and Camera-Trap Distance Sampling (CTDS) are two methods to estimate the population density of unmarked individuals (i.e., non-individually recognizable). However, these methods have not been previously evaluated together in low- density contexts, such as Scandinavia. Here, I assessed consistency and precision between the two models in six reference areas with five ungulate species in Sweden. I incorporated reported hunting statistics and local knowledge to determine the accuracy of the population density estimates. Further, I assessed the potential impact of extended sampling periods on the models where changes in animal activity are likely to occur due to behavioural and environmental shifts. Comparing REM and CTDS, I did not find significant differences in terms of density estimates in thirteen of sixteen populations. The REM was consistently more precise, with an average coefficient of variation of 0.29 compared to 0.43 for CTDS. Both models aligned with the independent proxies of density. Any potential impact of changes in behaviour influencing density estimates was masked by low model precision for both the REM and CTDS. Given that CTDS requires less effort for image processing, it is an attractive alternative to the REM, especially if the issue of low precision is addressed.