Leveraging diverse datasets to assess imperilled species declines: the Sunflower Sea Star in Canada as a case study

Evaluating the extinction risk of species is critical for establishing conservation priorities, yet efforts to do so are often hindered by a lack of targeted scientific monitoring data. Since myriad new forms of data are available from community-sourced (or citizen) science, new analytical approaches are needed to incorporate these data types into the analysis of population trends. Here, we describe an adaptable statistical framework that uses a combination of Bayesian hierarchical and state–space analysis to overcome the challenges of using disparate data types to derive a single estimate of population change. We apply this method to a status assessment of the Sunflower Sea Star ( Pycnopodia helianthoides) in Canada. By jointly analyzing quantitative observations from formal scientific surveys with semiquantitative observations from community-sourced surveys, we determined that Pycnopodia abundance in Canada declined by 82.1% (95% credible interval: 76.0%: 87.2%) following the onset of sea star wasting disease in 2013. Our approach substantially reduced uncertainty around the estimated decline relative to previous analytical frameworks and quadrupled the amount of data that could be used. Applying similar approaches to future species assessments will enable us to harness a greater diversity of data sources and reduce uncertainty in the conservation decision-making process.