Modelling complex habitat use for threatened bat species decision-making in landscapes with competing priorities

Species distribution models (SDMs) provide valuable information to aid conservation decisions, particularly in landscapes where economic and biodiversity priorities compete. Generating SDMs for species that rely on discrete habitat types for different activities (e.g. roosting or foraging) can be challenging, and result in outputs that are not appropriately tailored for end use. We collated expert-validated occurrence records for ghost bats (Macroderma gigas) and Pilbara leaf-nosed bats (Rhinonicteris aurantia), two threatened species in a region of intensive mineral extraction in north-western Australia. We generated spatial layers describing roosting and foraging habitat separately through literature review, expert consultation, and a novel neighborhood approach that inferred foraging habitat around roosts using summary metrics of key environmental predictors. Habitat suitability was then predicted using an ensemble SDM (averaging tuned-Maxent, Boosted Regression Trees, and Random Forest models). Through iterative consultation and co-design with end-users, outputs were refined into a spatial tool tailored for conservation decision-making. Roosting habitat for both species was largely predicted by moderately complex terrain and presence of major iron formations, as well as moderate terrain complexity and high vegetation diversity in the surrounding foraging neighborhood. Only 5–6% of predicted habitat occurred within conservation estate, while over 70% overlapped areas under mining tenure, with many known roosts located close to operational mines. Our approach demonstrates how explicitly modelling complex habitat use and co-designing SDMs with end-users can produce tools that better support landscape-scale habitat assessments across a region with competing priorities.