A Statistical Framework for Camera Trap Data Analysis in Ecological Research (R-9974)

Assessing anthropogenic causes of current animal density declines necessitates accurate quantification of population dynamics. Camera traps have become powerful, non-invasive, monitoring tools, but their inability to identify individuals of species without natural markings, in addition to limits in individual detectability, a well-known obstacle in faunistic abundance estimation, prevents the use of traditional analytical techniques. Recently proposed methods provide solutions, but often fail to account for multiple sources of bias and typically focus only on the analysis of a single target species. Moreover, it remains unclear how study protocols affect their performance and to what extent cost-effectiveness differs between study designs. Here, a statistical framework to analyze data from camera traps is proposed. My goal is to study mammalian population dynamics in the National Park Hoge Kempen (Belgium) and formulate conservation strategies, using data from a wide range of taxonomic groups, that I will collect myself, alongside experts and citizens. I plan to extend existing methodology into less assumption-driven models, for single species and communities, and to compare their potential, from statistical, practical, and economical perspectives. This will yield improved guidelines for data collectors and analysts, as well as a better understanding of mammalian abundances with the potential for direct implementation in conservation policies.

Keywords:BIODIVERSITY

Disciplines:Statistics, Landscape ecology