Site fidelity of harbor seals in Casco Bay, ME, USA using facial recognition technology: a pilot study

Abstract

Harbor seals, Phoca vitulina, play a critical role in regulating the biodiversity of coastal ecosystems in the North Pacific and Atlantic Oceans. We conducted a preliminary ecological study of harbor seals in Casco Bay, Maine using SealNet, a newly developed facial recognition software. We captured images of seals on nine haul-out sites to create a database of 768 seals in Middle Bay. We used photo ID techniques with facial recognition technology to record the location of individuals at each haul-out site. We calculated a range of 9% site fidelity to the Middle Bay inlet across years and 25% and 36% seasonal site fidelity to haul-out sites within 2020 and 2021, respectively. Preliminary estimates of the local seal abundance within Middle Bay ranged from 1562 (single haul-out site) to 2533 seals (across sites and years). In addition, our results suggest that the number of seals at haul-out sites is greatest from two hours before low tide to two hours after low tide and during high cloud cover conditions. We found no significant impacts of water or air temperature, level of boat traffic, or wind speed on haul-out site abundance. Our study supports the use of facial recognition technology as an effective method to monitor dynamic coastal species. The aim of future research will focus on a more systematic, longitudinal study design to monitor specific haul-out sites with the aim of providing more extensive connectivity data between sites and more refined estimates of site fidelity, turnover, and population size.