An investigative study of non-documentary television footage as a data source for community composition and biodiversity research

Summary Technology has created opportunities to explore non-traditional data sources for investigating patterns in community composition and biodiversity over time. Television shows set in natural environments implicitly contain ecological information. Through an investigative lens, we explored the potential utility of this data source for describing community composition and diversity trends given uncertainties in footage collection. We reviewed 14 seasons of the TV series Survivor (2016–2023) to quantify animal frequency and total appearances, describing the animals observed and calculating common diversity metrics. A total of 40 301 individuals representing 182 taxa were identified, with species accumulation analysis estimating up to 198 identifiable taxa detectable from Survivor footage. The number of animals observed was inherently biased towards more charismatic fauna, while the proportion of screen time was predominately given over to invertebrates. Temporal analyses suggested slight declines in biodiversity and species richness. However, we note several limitations that constrain the use of these data, particularly the unknown amount of sampling effort, as well as the absence of temporal and spatial information necessary for more robust analyses. Although we are currently sceptical of its application in ecological and conservation research, this data source may have greater value if paired with additional contextual information and metadata.