Breeding habitat selection of Canada Warblers across three distinct landscapes in Québec

We investigated the breeding habitat selection of the Canada Warbler (<em>Cardellina canadensis</em>) in southern Quebec across three distinct landscapes: a mixed boreal forest, an agroforestry landscape, and a riverine deciduous forest. This study aimed to characterize breeding habitat by assessing fine-scale vegetation structure and composition at surveyed stations during the breeding period. We conducted surveys between 2016 and 2018, with data collected on bird occurrence, vegetation structure, and plant community composition. We assessed habitat selection based on vegetation structure with logistic regression models and we tested associations with plant community using permutational ANOVAs and indicator species analysis. We observed variations in habitat selection across the three study areas. In the mixed boreal forest, Canada Warbler occurrence was positively associated with the density of large saplings and mixed-deciduous forest cover. In the agroforestry landscape, the species displayed a preference for mixed-coniferous forest cover and density of smaller saplings. In the riverine deciduous forest, no habitat variable was associated with the occurrence of the Canada Warbler as all surveyed stations were characterized by uniformly dense shrubby vegetation. Different plant species were indicators of Canada Warbler occurrence in each study area, but most of these species were part of the shrub structure. Although emphasizing a general preference for mixed forest compositions, our study highlights variations in Canada Warbler habitat selection across three distinct landscapes. Within predominantly coniferous landscapes, Canada Warbler demonstrated a tendency to select forest stands with a higher proportion of deciduous trees and, conversely, a higher proportion of coniferous trees within predominantly deciduous landscapes. Recovery and conservation efforts should focus on preserving or enhancing specific forest structures (i.e., high shrub density), while recognizing that measures may need to vary regionally.