UAV-based terrain metrics reveal species-specific nest-site preferences in Pygoscelis penguins

Abstract

Ongoing climate change and glacial retreat are expanding ice-free areas in Antarctica, offering new potential breeding habitats for marine birds such as penguins. This study examines the relationship between terrain features and the nesting distribution of three species of Pygoscelis penguins in the Western Antarctic Peninsula, a region experiencing rapid climate change, fishing pressures, and shifts in penguin species abundance and distribution. Here, we used remote sensing and geospatial analysis to process digital terrain models and orthophotos obtained from fixed-wing unmanned aerial vehicle (UAV) beyond visual line of sight (BVLOS) surveys (94.82 km and 104.20 km flights at altitudes of 350 m above sea level) in the South Shetland Islands, allowing detailed analysis of a selected 35-ha subset of the larger surveyed dataset. Statistical analyses confirmed selectivity in nest location and breeding ground choice by Adélie Pygoscelis adeliae, Gentoo Pygoscelis papua and Chinstrap Pygoscelis antarcticus Penguins. Adélie Penguin nesting areas differed from their surroundings in more relief variables than the nest-sites of Gentoo and Chinstrap Penguins. For all three species, at least one variable clearly distinguished breeding sites from the adjacent landscape. Adélie Penguins nested in higher, more inland areas, while Chinstrap Penguins occupied lower, coastal sites with steeper slopes and greater topographic variability. Gentoo Penguins used sites with the lowest hydrological risk, Chinstrap Penguins occupied sites with moderate water-accumulation conditions and Adélie Penguins nested in areas with the highest potential water accumulation, reflecting distinct ecological strategies across species. Our findings highlight preferred breeding conditions for each species, such as areas with minimal snow and water accumulation and a tendency toward low slope and northern exposure. These insights into nesting site selection could inform conservation efforts. The UAV BVLOS-based method demonstrated high efficiency and minimal disturbance, suggesting its suitability for further applications in penguin habitat monitoring and research.