Harnessing AI to Track Michigan's Wolf Pack

The vast forests and snowy landscapes of Michigan’s Upper Peninsula have long been home to a thriving population of gray wolves. However, accurately tracking the numbers of these elusive predators has proven to be a challenging task for wildlife researchers – until now. Thanks to the power of artificial intelligence (AI), a new approach is revolutionizing the way Michigan’s wolf population is monitored and managed.

The Michigan Department of Natural Resources (DNR) recently revealed the findings of its 2024 winter wolf population survey, which estimated a minimum of 762 wolves roaming the Upper Peninsula. This marks an increase of 131 animals compared to the 2022 estimate, further solidifying the region’s status as a stronghold for gray wolves.

This wolf survey unit stratification map shows varying degrees of wolf density across the Upper Peninsula. Image courtesy of MDNR

“This year’s survey findings are statistically consistent with our wolf population surveys for the past 14 years,” said Brian Roell, the DNR’s large carnivore specialist. “When a wild population reaches this stable point, it is typical to see slight variations from year to year, indicating that gray wolves may have reached their biological carrying capacity in the Upper Peninsula.”

Traditionally, wolf population surveys have relied on tracking footprints in the snow, a labor-intensive and time-consuming method. However, as Tyler Petroelje, a DNR wildlife researcher, explains, “The current minimum count requires significant effort to provide an index of abundance. As wolf density has increased, more time is needed to discern adjacent packs.”

Enter AI and a game-changing collaboration between the DNR and Michigan State University. Recognizing the limitations of traditional survey methods, the team embarked on a pilot project in 2022, deploying 200 trail cameras across portions of Marquette, Alger, Delta, and Schoolcraft counties.

One of the trail cameras – placed on tree trunks at a height of 4.5 feet from the ground – deployed for the 2023 wolf survey. Photo courtesy of MDNR

Over three months, these cameras captured an astonishing 1.7 million images, a treasure trove of data that would have been nearly impossible to sift through without the aid of AI. That’s where RECONN.AI, an artificial intelligence program, came into play.

“Breakthroughs in machine learning allow for rapid classification of remote camera images,” Petroelje said. “It also provides potential to monitor other wildlife species in addition to wolves.”

RECONN.AI proved its mettle, successfully identifying and sorting images of various animals, including 40,323 white-tailed deer, 7,534 black bears, and a remarkable 4,221 wolves. The software even went the extra mile, blurring images of people and vehicles to protect personal privacy.

This animation demonstrates how the RECONN.AI artificial intelligence software works to identify and sort various animal images from the survey photographs. Image courtesy of MDNR

Encouraged by the pilot project’s success, the DNR and Michigan State University team scaled up their efforts for the 2023-2025 survey period. This time, 1,230 trail cameras were strategically placed across 159 hexagonal cells covering the entirety of the Upper Peninsula, with one camera for every 16 square miles.

“Annual reports will be made available to the public,” Petroelje assured. “A public-facing website will also be created with interactive capabilities to view results.”

The implications of this AI-powered approach extend far beyond just wolf population monitoring. Roell noted, “The trail camera system will be less expensive because it will cut down on the amount of field time tracking labor. It can estimate the wolf population at other times of the year, allowing us to move away from a midwinter count.”

Wolf image gathered during the 2022 camera survey pilot project. Photo courtesy of MDNR

Moreover, the data collected by the cameras will provide valuable insights into the abundance of other wildlife species, such as bobcats, black bears, and moose, further solidifying Michigan’s commitment to preserving its rich natural heritage.

While the integration of AI into wildlife management is groundbreaking, it is not without its challenges. Roell acknowledged, “We have already heard about and witnessed a fair amount of camera sabotage damage,” underscoring the importance of public cooperation in ensuring the success of this innovative approach.

As the DNR and Michigan State University continue to refine and expand their AI-powered wolf monitoring program, the future looks promising for the preservation of Michigan’s iconic gray wolf population. With the aid of cutting-edge technology, researchers can now gather more accurate and comprehensive data, paving the way for informed management decisions and the potential delisting of gray wolves as a recovered species in the state.