Using citizen science image analysis to measure seabird phenology

Abstract

Developing standardized methodology to allow efficient and cost-effective ecological data collection, particularly at scale, is of critical importance for understanding species' declines. Remote camera networks can enable monitoring across large spatiotemporal scales and at relatively low researcher cost, but manually analysing images and extracting biologically meaningful data is time-consuming. Citizen science image analysis could reduce researcher workload and increase output from large datasets, while actively raising awareness of ecological and conservation issues. Nevertheless, testing the validity of citizen science data collection and the retention of volunteers is essential before integrating these approaches into long-term monitoring programmes. In this study, we used data from a Zooniverse citizen science project, Seabird Watch, to investigate changes in breeding timing of a globally declining seabird species, the Black-legged Kittiwake Rissa tridactyla. Time-lapse cameras collected >200 000 images between 2014 and 2023 across 11 locations covering the species' North Atlantic range (51.7°N–78.9°N), with over 35 000 citizen science volunteers ‘tagging’ adult and juvenile Kittiwakes in images. Most volunteers (81%) classified images for only a single day, and each volunteer classified a median of five images, suggesting that high volunteer recruitment rates are important for the project's continued success. We developed a standardized method to extract colony arrival and departure dates from citizen science annotations, which did not significantly differ from manual analysis by a researcher. We found that Kittiwake colony arrival was 2.6 days later and departure was 1.2 days later per 1° increase in latitude, which was consistent with expectations. Year-round monitoring also showed that Kittiwakes visited one of the lowest latitude colonies, Skellig Michael (51.8°N), during winter, whereas birds from a colony at similar latitude, Skomer Island (51.7°N), did not. Our integrated time-lapse camera and citizen science system offers a cost-effective means of measuring changes in colony attendance and subsequent breeding timing in response to environmental change in cliff-nesting seabirds. This study is of wide relevance to a broad range of species that could be monitored using time-lapse photography, increasing the geographical reach and international scope of ecological monitoring against a background of rapidly changing ecosystems and challenging funding landscapes.