Movement phases make a difference in habitat selection: Iberian lynx diversity of responses to human-modified landscapes.

Wildlife conservation requires a comprehensive understanding of habitat and landscape conditions supporting viable and well-connected populations, especially for reintroduction efforts. Habitat selection depends on different life processes and behavioural states driving animal movements. However, habitat modelling often considers all species records together or distinguishes only between locations within and outside home ranges. We considered a more detailed characterization of movements, focusing on five distinct movement phases: home ranges, transient residences, excursions, post-release dispersals and dispersals. Utilizing GPS telemetry data from 124 Iberian lynxes (Lynx pardinus), mainly tracked during a reintroduction programme, we examined their habitat selection and adaptation to heterogeneous, human-modified and fragmented landscapes. We developed context-specific mixed-effects habitat selection models for each phase, incorporating two hierarchical selection scales: resident range and path selection functions at the landscape scale, and step selection functions at the local scale. All lynxes consistently avoided intensive non-tree cropland and selected mosaics of natural vegetation, including tree, shrubland and grassland cover. Resident lynxes selected areas with low road and human infrastructure densities locally, but no such pattern emerged at the landscape scale, likely due to constraints on establishing large home ranges where infrastructures are widespread and intertwined with natural cover. We observed significant differences in habitat selection among non-residential phases. During excursions, lynxes avoided areas with high human infrastructure, whereas during dispersals, this avoidance was less pronounced, indicating high behavioural plasticity. The post-release dispersal phase mirrored dispersals but showed infrastructure avoidance and a higher selection for sheltering features like rugged terrain and shrub cover. Our findings emphasize the importance of distinguishing specific movement phases, and particularly different movement phases outside home ranges, when assessing habitat selection, especially for translocated animals settling in the landscape. This differentiation is essential to (i) identify suitable reintroduction areas with habitat features supporting post-release movements and settlement; (ii) evaluate habitat conditions of temporary stopovers facilitating long-distance dispersals, and (iii) achieve accurate connectivity estimates among populations. True dispersal events, as the primary movements facilitating gene flow and range expansion, should be treated separately in landscape permeability studies to effectively guide related conservation efforts.