An empirical assessment of the role of independence filters in temporal activity analyses using camera trapping data

Abstract

Independence filters are commonly applied to camera trapping data when determining the temporal activity patterns of species or populations. However, recent research challenges the validity of independence filters, stating that it leads to flawed inferences of temporal activity patterns and a substantial reduction in sample sizes, and should not be used. This study presents an empirical counter-argument using data from Snapshot Serengeti. By manually applying 60-minute independence filters, common behaviours influencing the reliability of activity patterns derived from density-based analyses in African savanna species were identified. Results showed that applying independence filters primarily leads to decreased midday activity densities, mainly due to herding behaviour and the prevalent use of shade in a tree-sparse savanna. Idle herds or individuals that feed or rest in front of camera traps also influenced derived activity patterns. Furthermore, it is confirmed that adequate sample sizes are generally maintained after applying independence filters. The findings from this study highlight the shortcomings of the recent research disputing the use of independence filters in temporal activity analyses and underscore the importance of a comprehensive approach of inspecting data, understanding species behaviour, and considering regional environmental characteristics. This will ensure more accurate representations of species’ true temporal activity patterns.