NORTHERN MICHIGAN – Researchers with the Michigan Department of Natural Resources will pilot trail cameras to estimate the state’s elk population. The effort is part of a broader study to identify the most accurate and cost-effective way to manage the herd.
The DNR’s Wildlife Division recently deployed about 200 cameras across the core elk range—roughly 1,100 square miles in northern Lower Michigan. Most are on state-managed land, though researchers also worked with private landowners.
Camera data will be compared to the DNR’s current aerial survey method to evaluate precision and cost.
Survey goals, challenges
Accurate population estimates help determine how many elk hunting licenses to issue annually. The DNR uses the harvest as its primary tool to maintain a sustainable population and offer recreational opportunities.
“We will compare the results from these different survey methods and balance the precision of the estimates they provide with the resources they require,” said Angela Kujawa, a DNR wildlife biologist.
For more than 20 years, the DNR has conducted aerial surveys every other year. The 2024 survey estimated 1,146 elk, with a confidence interval of plus/minus 262—meaning the population ranged from 884 to 1,408.
Tyler Petroelje, the DNR’s northern Michigan wildlife research specialist, said aerial surveys require reliable snow cover, which has been inconsistent during the January survey period. He noted the 25% uncertainty rate is acceptable for general wildlife management but less precise for Michigan’s target range of 500 to 900 elk.
Next survey, new technique
The next aerial survey is scheduled for 2026. Surveys typically take eight days, with two planes flying daily and two observers aboard each.
To test the new method, researchers will collect camera data each summer from 2025 to 2027. The battery-powered cameras, mounted about 40 inches off the ground, use infrared and motion detection to capture images stored on SD cards.
The DNR is partnering with the Michigan Cooperative Fish and Wildlife Research Unit at Michigan State University to analyze the photos using AI and human review:
- Images will be processed through MegaDetector, an open-source program that detects animals but doesn’t classify species. Human photos are deleted.
- Animal images will be uploaded to Wildlife Insights, a cloud-based platform that identifies elk. All uploads are reviewed by humans.
- MSU staff will verify elk identifications for accuracy.
Additionally, researchers will analyze teeth from harvested elk this fall and winter using statistical population reconstruction (SPR). This low-cost method supplements surveys but won’t replace them. If proven accurate, SPR could be used annually, allowing less frequent independent surveys—every three to five years instead of every other year.
Survey findings support Michigan’s elk management plan, which aims to maintain a sustainable population in balance with habitat.
“That management plan really guides the information we need to manage our elk herd,” Petroelje said. “And so we have a responsibility to follow the management plan to estimate elk abundance with the best tools that are available.”
