Are camera traps a reliable method for estimating activity patterns? A case study comparing technologies for estimating brown hyaena activity curves

• Published in: Remote sensing in ecology and conservation 2021-06, Vol.7 (2), p.129-138
• Main Authors: Edwards, Sarah, Noack, Jenny, Heyns, Louis, Rodenwoldt, Diethardt, Rowcliffe, Marcus, Hofmeester, Tim
• Format: Article
• Language: English
• Subjects: Activity pattern; Activity patterns; Animal behavior; brown hyaena; camera trap; Cameras; Collars; Estimation; Global positioning systems; GPS; GPS collar; Methods; Radio; radio‐tag; Sampling; Sampling methods; Sensors; Spatial data; Tags; Traps; Wildlife
• ISSN: 2056-3485; 2056-3485
• DOI: 10.1002/rse2.175

Camera traps and radio‐tags are both frequently and widely used sampling methods for deriving wildlife activity patterns. While radio‐tags continuously monitor a limited number of tagged individuals, camera traps have the potential to monitor all population members, albeit from spatially restricted, fixed points. Such differences might result in differing activity pattern estimates between the two sampling methods. However, few studies have compared the activity patterns derived from simultaneously employed sampling methods, or explored if using combinations of techniques might reveal greater insights into activity patterns. To address this, we compared the activity patterns derived from camera traps placed at latrines and den sites, both in combination and separately, to those from movement rates and the activity sensor collected by GPS collars on brown hyaena Parahyaena brunnea. The activity curve produced by combined camera traps showed a relatively high level of overlap with those produced by the movement rates and activity sensor; 0.88 (95% CI’s 0.87–0.89) and 0.85 (95% CI’s 0.83–0.86) respectively. However, camera traps reflected higher levels of activity from 00:00 to 06:00, than the radio‐tag methods, with data from den sites alone producing this higher level of activity. The results suggest that although hyaenas were active during the 00:00 to 06:00 period, they were so mainly at small, localized den sites, which reflected as lower levels of activity as derived by movement rates and activity sensors. Although our results reflect data from a single species and season, they illustrate the value of using a combination of techniques to disentangle complex behavioural activity patterns. Camera traps and radio‐tags are both frequently and widely used sampling methods for deducing wildlife activity patterns. We compared activity patterns derived from camera traps placed at latrines and den sites, both in combination and separately, to those from movement rates and activity sensors recorded by GPS collars on brown hyaena Parahyaena brunnea. The results suggest that although hyaenas were active during the 00:00 to 06:00 period, they were so mainly at small, localized den sites, which reflected as lower levels of activity than as derived by movement rates and activity sensors.