Using surveillance cameras to analyze the activity pattern of the Eurasian otters (Lutra lutra) and the efficiency of camera trap monitoring

Aim: The Eurasian otter (Lutra lutra) is a flagship species for global freshwater ecosystems. As one of the most widely distributed Palearctic mammals, the otter used to inhabit most provinces in China. However, after decades of relentless hunting and habitat degradation, the Sanjiangyuan Region is now one of last strongholds for this species in China. Currently, there still remains an enormous knowledge gap in basic information about this species. Available field information has mostly been acquired through camera trapping, and the reliability of this method is still being questioned considering the revealed underperformance in the monitoring for otters. Method: Research was conducted in Yushu City, Qinghai Province. We deployed surveillance cameras and camera traps in five monitoring sites with high otter occurrences from Oct. 2018 to May 2020. Using data collected from the surveillance cameras, we analyzed daily and annual activity patterns for the Eurasian otters in the region. To better understand the breeding period for otters, we analyzed the temporal distribution of all breeding related events. Furthermore, using surveillance cameras as reference, we evaluated the efficiency (detection rate, captured duration) and accuracy (behavior/topic, individual number) of camera traps for otter monitoring. Results: Surveillance cameras recorded 1,033 independent events of the Eurasian otters, and 597 of them were captured by camera traps (Oct.–May). The otters were highly active from 17:00 to 09:00 (+1 day) and there was a higher capture frequency from December to June (+1 year). Most breeding related events were captured from October to April (+1 year). Camera traps detected 69.18% of otter occurrence events, which was positively correlated with an increase in event duration (0–49 s). The temporal distribution of events captured by camera traps showed a significant linear relationship with surveillance camera records, but the durations were significantly shorter. Finally, only 56.28% of the events captured by camera traps provided enough information to discern the otter’s behavior; for events with more than one individual, only 49.35% of the camera trap records accurately captured otter’s number. Conclusions: The daily and annual activity patterns of the Eurasian otters in Yushu City were proven in accordance with the precious research conducted in other inland riverine ecosystems. By analyzing breeding related events, we found that the breeding season of the Eurasian otters in Yushu begins with mating behavior starting in October and ending with the dispersal of cubs by June. Although 30.82% of the events were missed by camera traps, they still accurately documented the daily and annual activity patterns of the Eurasian otters. However, because of the amount of missed information, camera traps are not the most reliable method for further quantitative behavioral studies on Eurasian otters.