River ice timing and duration in a warming climate: Field camera image classification

Timeline

2022

Description

This project is also part of Duke’s first Climate+ cohort.

Duke Data+ students, in collaboration with Dr. Emily Bernhardt (faculty advisor) and Audrey Thellman (graduate student) will evaluate how changing ice and snow conditions are impacting river ecosystems through classified ice imagery. Currently, our team has data from 7 field cameras that have been taking photos of the stream channel each day since 2018. We have created training data and code for a machine learning classifier to transform these photos into ecologically relevant indices, such as percent snow coverage. The Data+ team will modularize and visualize this classification pipeline to increase accessibility of our data product. Students will have the opportunity to work with a team of scientists at the New Hampshire site, U.S. Geological Survey partners with vested interest in the data product, and data scientists working in the Bernhardt Lab who have completed and are currently working on similar projects that increase availability and usability of environmental data (see https://cuahsi.shinyapps.io/macrosheds/).

Project Lead: Emily Bernhardt
Project Manager: Audrey Thellman, Weston Slaughter

View the team's final poster:

Image

Watch the team's final presentation below:

See what our students had to say about this 2022 summer project: