Occupancy dynamics of the Wood Warbler Phylloscopus sibilatrix assessed with habitat and remote sensing data

Abstract

The study of population dynamics and species–habitat interactions is important for many ecological and conservation issues. We analysed the occupancy dynamics of the Wood Warbler Phylloscopus sibilatrix, a forest interior passerine that has suffered long‐term declines in many European countries. Considering a broad range of habitat descriptors, including LiDAR‐based variables characterizing forest structure, we identified environmental factors influencing the probabilities of occupancy, colonization, extinction and detection, and mapped these probabilities at a landscape scale. We used presence–absence data collected between 1999 and 2013 in a robust‐design protocol with replicated surveys within each primary sampling season (i.e. breeding season). To analyse occupancy dynamics, we fitted dynamic occupancy models with covariates describing topography, forest extent and structure, soil, seed masting (as a proxy for rodent abundance) and weather conditions during settlement. Initial occupancy was positively related to slope, forest cover and the proportion of broadleaf forest, and negatively related to forest edge length, canopy height variation and soil nutrients. Colonization probability was positively associated with slope and negatively with canopy height variation and soil nutrients. Extinction probability was negatively related to slope, positively related to forest edge length and was lowest at intermediate vegetation heights. Detection probability decreased during the season and was positively related to forest cover. This study provides insights into the settlement dynamics of a declining forest species by identifying potential drivers of the processes underlying occupancy dynamics, instead of simply modelling species abundance or occurrence. Forest‐related covariates were of overwhelming importance for the settlement dynamics of the Wood Warbler. Conservation measures should focus on providing and maintaining the species' preferred forest structure, specifically in extended forests. In Switzerland, forests located on slopes and on nutrient‐poor soils should be prioritized for conservation. Finally, comparing findings derived from the relatively coarse scale of this study with those from the small scale of individual territories suggests that conservation recommendations deduced from large‐scale monitoring schemes alone may be incomplete, and vice versa.