Camera traps reveal seasonal variation in activity and occupancy of the Alpine mountain hare Lepus timidus varronis

Abstract

Mountain hare is a cold‐adapted species threatened by climate change, but despite its emblematic nature, our understanding of the causes of population decline remains limited. Camera traps are increasingly used in ecology as a tool for monitoring animal populations at large spatial and temporal scales. In mountain environments where field work is constrained by difficult access and harsh conditions, camera traps constitute a promising tool for surveying rare and elusive species such as the mountain hare. Our study explored the use of camera traps as a tool for studying seasonal habitat occupancy and daily activity patterns of the mountain hare, in order to carry out long‐term monitoring of populations. We installed 46 camera traps along elevation gradients in the Mont‐Blanc massif (France) from January 2018 to June 2022. We measured habitat variables at each camera trap site in order to define vegetation composition and habitat structure. We performed multi‐season and single‐season occupancy models to respectively describe habitat occupancy of the mountain hare throughout the year and identify the environmental variables influencing mountain hare presence during the breeding season. Mountain hares occupy coniferous forest in winter, and then switch to mixed areas of shrubland and grassland above treeline in spring and the beginning of summer. In spring, occupancy probability of the mountain hare increases with relative cover of mixed low shrub and herbaceous layer (i.e. the 10–40 cm vegetation layer), suggesting a link to food resources and protection from predation. Our results also confirm the nocturnal and crepuscular activity of the mountain hare during the breeding season, and strictly nocturnal activity in winter. Our results demonstrate the efficiency of camera traps as tools for monitoring mountain hare habitat occupancy in mountain environments and underline the importance of diverse habitat mosaics for the preservation of the species.

Zero-shot animal behavior classification with vision-language foundation models

“…Therefore, this approach could enable large-scale annotation of camera trap datasets and facilitate or automate the study of behaviors, which could for example be particularly useful for studying the effects of climate change or human activities on the behavior of wild animals. These research topics are increasingly studied using camera-traps (Bison et al 2024;Blount et al 2024). Furthermore, even though this study illustrates the effectiveness of foundation vision-language models for a specific task, the flexibility of the method makes it easily adaptable to other ecological tasks.…”