Date of Award

8-17-2024

Document Type

Thesis

Abstract

Ravens, Corvus corax, and other corvids are intelligent birds that are the focus of many studies, such as in-depth dives into potential facial recognition and tool use to name a few. Despite these numerous behavioral studies, ravens lack an accessible basic universal ethogram and have rarely been observed in their undisturbed, natural state. Due to this, my study focuses on free-roaming common raven behavior in Fairbanks, Alaska, for which I utilize exploratory analysis to identify patterns in collected data. In doing so, I show how data mining and machine learning can further support behavior research with a systems perspective in the Anthropocene using pattern recognition. Using an ethogram and machine learning techniques on open access data for two winter seasons, I examine what factors affect common raven behavior around human-subsidized food sources in Fairbanks, Alaska by answering: 1) What consistent reactions do wild ravens communities show to objects, people, and other organisms (typically small songbirds or dogs) and 2) Do other factors, such as daylight or location, contribute to differing raven behaviors? I found that ravens exhibit predictable responses that vary based on urbanization level. In addition, I found an unusual pattern in raven behavior that indicates that ravens adjust their behavior based on hourly and daily human activity, indicating that raven behavior is scheduled. These results provide evidence that merging modern and classic techniques into behavioral research reveals patterns that may be missed by traditional methods alone.

Handle

http://hdl.handle.net/11122/15454

Share

COinS