A Bayesian approach to include Indigenous Knowledge and satellite telemetry data in habitat selection functions

Indigenous Knowledge (IK) of species‐habitat relationships provides in‐depth understanding of species presence. While this is the case, IK and ‘Western’ scientific data are rarely combined in the same modelling approaches and species management remains mostly within colonial frameworks with limited capacity to include different data sources, such as IK, in management decisions. In partnership with Iñupiat hunters, we developed a method to include IK in Bayesian habitat selection functions as priors and show how habitat variables mapped by IK holders can be included as covariates, with satellite telemetry data. We applied our methods in a case study for ringed seals ( natchiq in Iñupiaq; Pusa hispida ) in habitat selection function models that also include dynamic habitat types (e.g. ice concentration). We show, through our case study, that IK and Western scientific data can reflect similar ecological processes and how IK can provide extended insight into broad habitat use that are not captured by Western methods alone. We demonstrate that when spatial IK is included as new categorical variables it allows for the inclusion of environmental covariates that may not be available from other data sources at an appropriate scale and can account for complex habitat types that would not be possible to model otherwise. Through our case study, we show the importance of the inclusion of IK to provide fully informed models predicting spatial variation of ringed seal relative probability of use that can be used to inform critical habitat designations. Synthesis and applications . Our work points to the importance of both Western science and IK data sources in models used for conservation and management decisions. We show how the models that include the IK result in more informed predictions and the value of IK when predicting important areas for species that may be relied upon for conservation and management decisions. Our results highlight the impact that the inclusion of IK can have in modelling outputs, with IK providing needed insights into the potential limitations of Western scientific data, leading to a more informed understanding of species‐habitat use.