Date of Award

Fall 12-16-2022

Level of Access Assigned by Author

Open-Access Thesis

Degree Name

Master of Science (MS)

Department

Ecology and Environmental Sciences

Advisor

Cynthia Loftin

Second Committee Member

Daniel Hayes

Third Committee Member

Erik Blomberg

Additional Committee Members

Linda Welch

Glen Mittelhauser

Abstract

Seabirds have been identified as one of the most threatened groups of birds. Declines in their populations have been attributed to a variety of factors, including overfishing, invasive species, environmental pollutants, and climate change to name a few. Additionally, some seabird species come into conflict with humans and other wildlife, creating additional need for long term monitoring of seabird populations regardless of past or present ecological success. Understanding the distributional patterns and trends of these aggregations of birds has intrigued researchers, as seabirds have been regarded as “indicators” of the health of marine ecosystems owing to their susceptibility to environmental changes.

Aerial surveys have been used to survey wildlife in a variety of contexts, as ground surveys can often be logistically difficult, time consuming, expensive, as well as disruptive to colonies. Counts derived from aerial imagery may be less subject to observer bias than real-time surveys, however variation in detection and subsequent interpretation or classification can introduce observer bias objects.

We first determined the components that affected whether an observer failed to detect and subsequently, properly classify, seabirds belonging to six species using a multi-observer approach. We used multinomial and logistic regression models separated by species to determine whether or not likelihood of interpretation errors made by observers varied by species, behavior, and spectral context and heterogeneity. We found that detectability of birds in aerial imagery varied by species in particular, with some species such as terns (Sterna sp.) being more likely to be missed by observers as compared to other species. We also found heterogeneity in detection and misclassification probabilities depending on the habitat type in which individual birds were located.

Second, we determined the components of imagery that affected whether an observer made detection errors and behavioral misclassification errors. We then used the interpretations of observers to derive coast-wide counts of the number of gulls and Double-crested Cormorants (Phalacrocorax auritus) nesting on Maine’s coastal islands in 2019 and evaluate changes in the distribution and counts of these seabirds from 2008 to 2019 on Maine’s coastal islands. Great Black-backed Gulls (Larus marinus) experienced great declines across the coast of Maine between 2008 and 2019 while Herring Gulls (Larus argentatus) and Double-crested Cormorants leveled off between 2013 and 2019 despite previous declines in previous survey efforts. Our work not only provides insight on how to improve future seabird survey efforts by accommodating for heterogeneity in detection and misclassification probability, but also identifies populations in decline that may require additional research efforts.

Download

Share