Long-Term Camera Trap Data and Community Structure Analysis

Abstract

Long-term camera trap monitoring provides an invaluable window into mammal community dynamics, occupancy trends, and interspecific interaction networks that are inaccessible to short-duration surveys. This study analysed 11 years of standardised camera trap data (2010–2021) from a 2,400 km2 mixed forest–agricultural landscape network across France, Sweden, and Germany (48 permanent stations; 142,680 trap-nights; 284,642 independent detection events for 34 mammal species). Community structure was characterised annually using occupancy modelling, temporal activity overlap (Dhat 4 ), and co-occurrence analysis (C-score, B-score). Species richness showed a significant increasing trend over the monitoring period (Poisson GLM: b = +0.028 per year, p = 0.002), driven primarily by range expansion of Capreolus capreolus, Sus scrofa, Lynx lynx, and Canis lupus. Apex predator (wolf and lynx) occupancy increased significantly after 2014 (before-after comparison: occupancy 0.18 to 0.48; z = 4.82, p < 0.001), coinciding with a significant decline in roe deer nocturnal activity overlap with cameras (Dhat 4 = 0.82 to 0.64), consistent with a landscape-scale anti-predator behavioural response. Co-occurrence analysis revealed significant positive associations between apex predators and large ungulates (B-score: 0.84 ± 0.08), and significant negative associations between mesopredators (red fox, stone marten) and apex predators (C-score: -0.72 ± 0.06), consistent with mesopredator release suppression by apex predators. Temporal activity overlap between competing species declined significantly over the monitoring period as community complexity increased, suggesting niche partitioning through temporal segregation. These findings demonstrate the value of long-term camera trap networks for detecting community-level responses to apex predator recovery and provide a methodological framework for cost-effective large-scale biodiversity monitoring.