Integrating remote sensing and species' traits to assess bird responses to wildfire in agropastoral landscapes

The climate-induced increase in wildfires has made them a critical research priority for biodiversity in Europe. Yet, studies leveraging species traits to explain varying responses to wildfires and inform targeted conservation efforts remain limited. To address this gap, we surveyed bird communities one-year post-wildfire across a 4450-hectare burnt and adjacent unburnt agropastoral landscape in the Classic Karst, a protected area spanning Italy and Slovenia undergoing reductions in agropastoral farming. Remote sensing data and machine learning were employed to map wildfire severity and finely classify bird-habitat associations based on 139 bird point counts. Wildfire substantially reduced woody biomass and was associated with lower bird species richness (mean/point = 5.09 SE ± 0.22) and abundance (mean/point = 6.69 SE ± 0.35) compared to unburnt areas (richness mean/point = 6.43 SE ± 0.22; abundance mean/point = 8.71 SE ± 0.32). However, its impacts varied considerably depending on species traits. Among the species with the lowest post-fire occurrences were high shrub and forest-dwelling birds that rely on vegetation cover, particularly those with open-cup nests and branch-edge foraging behaviours. Conversely, cavity nesters and trunk/bark foragers were more frequent in burned areas, potentially due to the structural resilience of their nests and the post-fire availability of saproxylic invertebrates. Some semi-open habitat species of conservation concern, particularly those employing ground-nesting and ground-foraging strategies, also showed increased occurrence in burnt areas. Smaller-bodied species were more frequently detected in heterogeneous unburned shrub patches, consistent with the Textural Discontinuity Hypothesis. Resident/short-distance migrants appeared less affected than long-distance migrants, probably due to their continuous or partial presence year-round. Our results highlight the value of combining trait-based and remote sensing analysis to identifying species responses following wildfire, providing critical insights for conservation strategies that aim to balance the needs of both open-habitat and forest species in Mediterranean fire-prone landscapes.