'Fire birds' freeze for the camera

Authors

Joshua Dale

Document Type

Article

Publication Date

11-5-2012

Abstract

'Fire birds' freeze for the camera

Human activity impacts wildlife ecology, and Susan Ellis-Felege is hunting down the details.

Three years ago, North Dakota Game and Fish (NDGF) began a five-year study on demographic parameters in and around energy development areas in Western Dakota. They wanted to know what effects, if any, harvesting of natural resources has on the survival, movement, and reproduction of wildlife.

When Felege, known for her game bird research at the University of Georgia, migrated to UND in 2011, the NDGF jumped at the opportunity to add her unique expertise to the mix.

"I felt honored," said Felege. "I am an applied wildlife ecologist and any time I can work with agencies, it provides me with an opportunity to use my expertise to inform management decisions. NDGF has a long history of working with UND and I am excited to have the opportunity to continue that tradition."

Predators, too

Specifically, Felege is studying the nesting habitat of sharp-tailed grouse — known as "fire birds" by American Indians — in the Belden, Blaisdell, and Lostwood Refuges around Tioga, N.D., in western North Dakota. Belden has the most human activity, followed by Blaisdell and Lostwood.

Felege also studies predator interaction with the birds.

"In ecosystems, everything interacts," Felege noted. "Studying or focusing on just one element provides only a limited viewpoint of interactions occurring. More recent approaches in wildlife management and conservation involve managing ecosystems and not just a single species."

Marriage of biology and technology

To monitor the nesting habitat, Felege uses a digital video camera smaller than a pop can. She is capable of filming continually for about four days using modern digital and LED technology. An inconspicuous cable connects to a waterproof case that houses a miniature computer recording 32 gigabytes of data at a time — equivalent to 20,000 digital pictures. Motion and heat activated cameras are not used because of their potential fail rate in the dense grasses.

"They are a big improvement over the last cameras," said Felege. "They were only able to last a day before we had to change a marine battery and VHS cassette."

This year's research has captured more than 30,000 hours of footage. Felege is projecting more than 50,000 hours of video will be taken over the course of her project. And most of the footage will be hours of inactivity.

To help her store and analyze this data, Felege turned to UND computer scientist Travis Desell and his experience with "Distributive Computing" to search for the segments of interest.

Grouse are designed to blend well into their environment. And while it's a strain for the human eye to cut through the grouse's camouflage, it's even more challenging for computers, which largely rely on color contrast. "Computer vision," a computer's ability to distinguish objects in an image, is still in its primitive stage. Desell must actually teach or train the computers to recognize differences between vegetation and animals and their movement.

"Basically, we need to separate movement of the grouse and predators from wind blowing the grass around using computer-vision techniques," Desell said.

Unlike humans, computers don't fatigue after hours of video analysis and are able to accomplish the task quickly and efficiently, allowing researchers to focus more on the subject matter.

It's still too soon to say what effect harvesting of energy resources and natural predation is having on sharp-tailed grouse in western North Dakota. But whatever is discovered, Felege says, it won't necessarily be all negative.

"We hope to learn about potential positive and negative impacts of gas and oil on grouse, predators, and interactions of these animals that will help wildlife biologists make informed management decisions about these species in areas where gas and oil development is occurring."

This document is currently not available here.

Share

COinS