Habitat and predator heterogeneity influence density of a declining mammal

Accurate density estimates are crucial for effective conservation management. However, in highly dynamic landscapes where variation in habitat composition and predator–prey interactions in both space and time is likely, integrating spatiotemporal covariate effects in density estimation is challenging and often large datasets are needed. Here, we used an 11‐year spatial capture–recapture (SCR) dataset from a typical mixed agroecosystem in England to estimate landscape‐scale densities of western European hedgehogs Erinaceus europaeus. We simultaneously integrated spatially varied habitat covariates, and the spatiotemporal variation in predator (Eurasian badger Meles meles) den site into one SCR framework. Density was spatially structured (range 0.39–13.54 on a 1 km2 grid), and was lower in arable fields and highest in amenity grasslands next to buildings. Density was also positively associated with soil permeability, density of edge habitats, proximity to the nearest building, and distance from the nearest badger sett. A new badger sett appeared halfway through the study period, resulting in a hedgehog density‐weighted population centre over the study area shift away from the badger sett and a decrease in annual hedgehog density estimates, supporting the landscape of fear for hedgehogs in response to their main predator the badger. Density estimates were also 43% lower after incorporating spatiotemporal covariate heterogeneity into the modelling process, highlighting the need to integrate dynamic habitat and predator influences into density modelling to provide more accurate estimations. Finally, our findings demonstrate the importance of long‐term monitoring for understanding population responses to changes in predator presence and provide clear empirical evidence for a prey species altering space use in relation to the increased predator, supporting the landscape of fear hypothesis.