Combining scales in habitat models to improve conservation planning in an endangered vulture

Predictive modelling of species’ distributions has been successfully applied in conservation ecology, but effective conservation requires predictive and accurate models. The combination of different scales to build habitat models might improve their predictive ability and hence their usefulness for conservation, but this approach has rarely been evaluated. We developed habitat-occupancy models combining scales from nest-site to landscape for a key population at the northwestern edge of the distribution of the globally endangered Egyptian vulture (Neophron percnopterus). We used generalised linear models (GLM) and an information-theoretic approach to identify the best combination of scales and resolutions for explaining occurrence. Those models that combined nest-site and landscape scales improved the predictive ability compared with the scale-specific ones. The best combined model had a very high predictive ability when used against an independent dataset (92% correct classifications). Egyptian vultures preferred to nest in caves with vegetation at the entrance that were situated at the base of long cliffs, provided that these cliffs are embedded within low-lying, heterogeneous areas with little topographic irregularity and with little human disturbance. The density of sheep around the nest positively influenced Egyptian vulture presence. Conservation of the studied population should focus on minimising human disturbance and on promoting sustainable development through conservation of traditional pastoralism. Our findings highlight the importance of developing region-specific multiscale models in order to design effective conservation strategies. The approach described here may be applied similarly in other populations and species.