Habitat mediates coevolved but not novel species interactions

  • Joshua Twining (Contributor)
  • Chris Sutherland (Contributor)
  • Neil Reid (Contributor)
  • David Tosh (Contributor)



On-going recovery of native predators has the potential to alter species interactions, with community and ecosystem wide implications. We estimated co-occurrence of three species of conservation and management interest from a multi-species citizen science camera trap survey. We demonstrate fundamental differences in novel and co-evolved predator-prey interactions that are mediated by habitat. Specifically, we demonstrate that anthropogenic habitat modification had no influence on the expansion of the recovering native pine marten in Ireland, nor does it affect the predator’s suppressive influence on an invasive prey species, the grey squirrel. In contrast, the direction of the interaction between the pine marten and a native prey species, the red squirrel, is dependent on habitat. Pine martens had a positive influence on red squirrel occurrence at a landscape scale, especially in native broadleaved woodlands. However, in areas dominated by non-native conifer plantations, the pine marten reduced red squirrel occurrence. These findings suggest that following the recovery of a native predator, the benefits of competitive release are spatially structured and habitat specific. The potential for past and future landscape modification to alter established interactions between predators and prey has global implications in the context of the on-going recovery of predator populations in human-modified landscapes.,A survey spanning a five-year period documenting the occurrence of pine marten and grey and red squirrel was conducted throughout Northern Ireland between 2015 and 2020. The survey was repeated three times, initially in 2015 with 332 sites surveyed by citizen scientists provided with camera traps and trained for their consistent use (for full Methodology see Twining et al. 2021). This survey was repeated in 2018 with 172 sites, and in 2020 with 207 sites using the same methods. At each site, a single camera trap was deployed at a point randomly selected by the surveyor within an independent 1km grid. Cameras were installed at head height on a tree overlooking a wooden squirrel feeder erected on an adjacent tree. Feeders were baited with peanuts and sunflower seeds in 2015 and 2020, but just sunflower seeds in 2018. Cameras were set to take three images per trigger with a 1s reset time. Camera traps were deployed for 7 - 14 days at each location (mean = 10.3 days) after which cameras were retrieved for data extraction and species identification. Detection records were created for each species over the recording period.,Dates not sampled are replaced with NA's Each row the detection - non detection .csvs aligns with the same line on the associated site level covariate file,
Date made available14 Dec 2021

Cite this