Identifying wild carnivore behaviors from accelerometer signatures

Department for the Ecology of Animal Societies

Project description
The student will work with tri-axial accelerometer data to characterize a suite of common movement behaviors (e.g. climbing, walking, jumping, running) using manually labeled video recordings of captive fosa Cryptoprocta ferox, the apex predator of Madagascar. The primary objective of this project focuses on how sampling regime of accelerometry data alters the ability to identify common animal behaviors. Wild animal accelerometer data are often recorded at a reduced rate or duration to compensate for memory and battery constraints of the devices. To test this, the student will use a supervised learning approach to identify different acceleration patterns down sampled at different rates and intervals to evaluate their accuracy to identify a subset of the most clear and common accelerometer signatures in wild fosa data collected from Madagascar in 2022-2023. Collecting data using captive animals allows us to study aspects of fosa behavior and biology that would be almost impossible to obtain in the wild due to their elusive nature. Being able to transfer this knowledge to wild fosa, can ultimately provide us with a better understanding of wild animal responses, movements, energetic budgets, seasonal behavior, etc. Additional research questions may be developed depending on the student’s interests. For instance, we are also interested in broadly examining energetics via VeDBA (vector of dynamic body acceleration) metrics for wild collared fosa to classify active vs. passive behavioral periods in connection with seasonal movement patterns.

Preferences

  • BA in ecology, biology, computer science or a related field

  • Coding experience using R or a similar computing language

  • Interest in behavioral and animal ecology

  • Organized, self-motivated and able to work independently

  • Competency writing and communicating in English

  • Prior experience with accelerometer data or animal movement data


Project Supervisors
Maeva Ratsimbazafindranahaka, Zea Walton, Richard Gunner, and Meg Crofoot, Department for the Ecology of Animal Societies

How to apply
Please send a short motivation letter and CV to: Maeva Ratsimbazafindranahaka, or Zea Walton

Go to Editor View