Large carnivores influence ecosystem dynamics in multiple ways, e.g. by suppressing meso-carnivores and providing carrions for smaller scavengers. Loss of large carnivores is suggested to cause meso-carnivore increase and expansion. Moreover, competition between meso-carnivores may be modified by the presence of larger carnivores. In tundra ecosystems, the smallest meso-carnivore, the Arctic fox, has experienced regional declines, whereas its larger and competitively superior congener, the red fox, has increased, potentially due to changes in the abundance of apex predators. We explored if variation in occurrence of wolverine and golden eagle impacted the occurrence and co-occurrence of Arctic fox and red fox in relation to varying abundances of small rodents within the Scandinavian tundra. We applied multi-species occupancy models to an extensive wildlife camera dataset from 2011-2020 covering 98 sites. Daily detection/non-detection of each species per camera trap site and study period (late winter; March-May) was stacked across years and species occupancy was related to small rodent abundance while accounting for time of year and status of simulated carcass. Arctic fox was more likely to co-occur with red fox when wolverine was present, and less likely to co-occur with red fox when golden eagles were present and wolverine absent. Red foxes increased in occupancy when co-occurring with the larger predators. Arctic fox responded more strongly to small rodent abundance than red fox and co-occurred more often with the other species at carcasses when rodent abundance was low. Our findings suggest that the interspecific interactions within this tundra predator guild appear to be surprisingly intricate, driven by facets of fear of predation, interspecific mediation and facilitation, and food resource dynamics. These dynamics of intraguild interactions may dictate where and when conservation actions targeted towards the Arctic fox should be implemented.

Data was collected using a network of wildlife cameras in four alpine and one Arctic region of Norway in the period 2011-2020. Cameras were active during late winter each year (March-May) and was baited with scraps and trimmings from from slaughtered reindeer (15-20 kg), simulating a carcass. Cameras were checked and baits replaced once after approximately three weeks. The cameras were set to a time-lapse interval of 5 minutes, and the final data set was aggregated into daily presence/absence of each species.

Files include the daily aggregated detection data on the four focus species (Arctic fox, red fox, wolverine and golden eagle) in CSV format. The data is split into two regions; alpine and Arctic, which are modelled separately. The R script contains routines for data import and modelling. See README.md for further details.