University of Michigan School of Information
UMSI students advance wildlife research with motion-sensor camera data
Tuesday, 07/16/2024
A pair of desert bighorn sheep lean their horns together and stare directly into a camera lens, with the dusty red rocks of Utah’s Capitol Reef National Park as their backdrop. They look like they’re taking a selfie.
Motion-sensor wildlife cameras — which have been installed throughout many U.S. national parks — offer close-up glimpses of wildlife, as we’re rarely able to see them: Desert bighorn sheep butt heads. Mountain lions roam in the dark. A porcupine clambers up a rock with quills that look deceptively soft.
But this footage isn’t purely for entertainment. Wildlife cameras enable the National Park Service to track the movement and behavior of animals, monitor the growth and decline of species and even mitigate disease.
“The wildlife monitoring project is noninvasively studying species that call Capitol Reef home in order to encourage and preserve biodiversity,” explains Brooke Rossow, a Bachelor of Science in Information student on the information analysis pathway at the University of Michigan School of Information.
All BSI students complete an advanced project course in their final year, which allows them to put their training into practice. Rossow and her team members Harrison Barth, Abby Finn and Miles Scheffler used their skills in data wrangling to facilitate wildlife monitoring in Capitol Reef National Park.
They developed a script to automate the reporting and analysis of photo data from wildlife cameras, along with visualizations to help park managers draw insights from existing data.
“With all this camera data that they have, they didn’t really have a good means of quickly categorizing it and reporting findings, so it was sitting for a long time,” says Finn.
She is the only member of the team who has visited Capitol Reef, which she describes as “a mix of all the best parks in Utah,” with its striking arches and striated rocks. But all four students felt a personal tie to the project.
“We chose this project because we feel some connection to nature or some desire to preserve nature,” Barth says. “Whether we're talking about controlling wildlife, preventing disease or reinhabiting areas, our project would make that easier for park management. It will require less manpower and hours. They'll be able to do these things they're trying to do at a greater scale.”
The National Park Service is a returning client for UMSI, having hosted seven course projects in the past five years. To protect the national parks and educate the public about them, the NPS relies on a resource UMSI students can provide: information.
The case of the disappearing, reappearing sheep
Desert bighorn sheep are a key focus for park managers and ecologists because they are considered sensitive to extinction. Across the West, the species began to decline in the late 1800s because of disease, habitat loss and unregulated hunting, until they were eradicated from Capitol Reef in the 1940s.
A few decades ago, desert bighorn sheep were reintroduced to the park. But they still face many threats. One of these threats might seem surprising.
“Pneumonia, within this population, spreads incredibly quickly and widely,” Finn says. “So, being able to track their migration patterns and the presence of signs of disease is really important to monitoring the health of the species.”
Wildlife cameras can capture telltale signs of pneumonia outbreaks in desert bighorn sheep, including foaming at the mouth.
On the UMSI team, Finn and Scheffler analyzed and visually represented data related to desert bighorn sheep. For each camera capture, they worked with a set of metrics that included the time stamp, region of the park, sex of the animal, and signs of disease.
One of their graphs shows the annual percentage of desert bighorn sheep sightings in Capitol Reef broken down by sex (male, female or unknown). Another represents the regional detection of desert bighorn sheep by age.
Meanwhile, Barth analyzed temporal data to see how often certain species were detected in different park regions throughout the year. He worked to create “a historical roadmap of how species have progressed over time and space.”
Ask any member of the team about the challenges this project presented, and they’ll say without hesitation: dealing with repetitions. This is the central problem in the movie “Groundhog Day,” too. But here it means something different.
“For example, you have a sheep that comes in front of the camera and it takes a photo of him, and then he comes back 30 seconds later and it takes another photo. The camera can’t really detect that that’s a repeat,” Barth says. “We had to find our own ways to deal with discrepancies in the data.” This might mean, for example, only counting one detection per species in a given day.
Another challenge: To meet the needs of the park, the team learned to use a tool that park managers use — not a radio or compass, but a programming language called RStudio.
“I was very impressed that the students were willing to learn a new coding language, which was specifically requested since park managers have some experience in R,” says Joseph Ceradini, an ecologist at Capitol Reef who worked closely with the student team.
“Especially going into a post-grad job where there's probably going to be new software you have to learn, I feel like this has just shown me how to be able to work through that on my own,” Rossow says.
The students have brought Capitol Reef National Park’s wildlife photo database to life.
The team’s final product is a dynamic report in R that visualizes trends across years of camera data. It features interactive graphs that answer 20 important questions from park management related to the migration patterns of desert bighorn sheep and other species, regional hotspots in the park, and critical cameras.
“The students have brought Capitol Reef National Park’s wildlife photo database to life,” Ceradini says. “Prior to their involvement, we had a lot of cool pictures of wildlife but no formal summaries of the wildlife photo data, which means it was not possible to reliably use the photo data to help answer management questions. Now, we can run the script the students developed with new data any time and have up-to-date summaries of the wildlife photos.”
According to Ceradini, wildlife monitoring using motion-sensor cameras is becoming increasingly common — and increasingly valuable. This technology allows park managers and wildlife researchers to collect more data than ever before, and to do so noninvasively.
“This script will be a useful tool for park managers for many years to come,” he says.
The goal of any UMSI project is sustainability — students aim to create something that can be used by their client organization for a long time. But working with a national park brings another layer of meaning to “sustainability.”
The team hopes their project will help to ensure that the diverse species in Capitol Reef National Park are protected now and into the future. So yes, keep the cute photos of desert bighorn sheep coming.
—Abigail McFee, marketing and communications writer;
video by Jeffrey Smith, multimedia producer
LEARN MORE
Client-based courses provide UMSI students with high-impact learning experiences. Students gain confidence communicating with a client, while developing deliverables they can include in their portfolios. Learn more.
Organizations from all industries and sectors are invited to propose information-based projects for students to work on through client-based courses, capstone projects and community-based programs. Host a student project.