My MSc is about understanding the spatial and temporal variation in habitats selected by rare species and how this is influenced by our choice of modelling approach. My primary case study is the Yellow Rail (Coturnicops noveboracensis), a secretive wetland bird, for which we began passive acoustic monitoring in 2012. We have developed automated recognizers to detect and classify the vocalizations of Yellow Rail and other species targeted in the bioacoustics unit’s rare animals monitoring program in the Lower Athabasca. Patch-occupancy models are often used to evaluate species distributions. I will be comparing their outcomes when data from our bird experts is used vs. data from our recognizers. Additionally, these models explicitly correct for detection error but I’d like to explore how the relative importance of this changes when you survey more intensively.