Churchill, CanadaPolar bears, the world’s largest land carnivores, are spending more time ashore in search of food as climate change melts their prime hunting ground—sea ice. Increasingly, this results in bears entering Arctic communities, where they sometimes come face to face with residents. In February 2019, a “mass invasion” on the Russian Arctic archipelago of Novaya Zemlya trapped people in their homes while several dozen bears wandered the town. In 2018, two separate attacks in the Canadian province of Nunavut resulted in the deaths of two men; in 2021 a bear attacked and injured three people in the same region before it was killed.
But an existing technology modified to spot the white, fluffy bears may be able to help reduce that threat: portable radar units to warn communities that a bear is in the vicinity. Polar Bears International (PBI), an organization that supports research on and conservation of polar bears and sea ice, has been at the forefront of developing and testing such systems—nicknamed bear-dar—for the last several years, and envisages a number of scenarios in which they could be of use.
For example, in parts of their range polar bears are increasingly forced to scavenge from trash dumps. PBI’s Senior Director of Conservation Geoff York notes that in northern communities where securing the landfill is a significant challenge, a bear-dar could alert people that a bear is in the dump. “So, if you’re going to drop off your trash, or if you’re going to work, you know at the very least to take extra care,” he says.
PBI and its collaborators have been testing several different prototypes, all of which involve a combination of a radar, a camera, and either a WiFi or cellular network to transmit the images and data to a server and then to a computer or smartphone. One transportable version, which is mounted on a collapsible tower on a trailer, has been in operation for several years on the outskirts of Churchill, Manitoba, the “polar bear capital of the world.” The organization has recently begun trialing a much smaller device, developed by students from Brigham Young University, which it has mounted on the specially designed “buggy” it uses to study and film polar bears outside Churchill.
“It's a brilliant little package where they have incorporated the radar, a little processor, and a camera all in one unit in a little gray plastic box about the size of an iPad,” explains B.J. Kirschhoffer, PBI’s Director of Conservation Technology.
How bear-dar works
The idea behind all the systems is simple. The radar detects an object, at which point software determines whether that object is a bear and, if so, sends out an alert to whoever a community designates. That person can pull up video from the camera in real time on their phone and check it. “And then they can decide: ‘Is this something I need to take action on? Or can I go back to bed? Is it a bear, or is it just an Arctic hare or a caribou?’” explains York. Or it could send an alert to a community’s social media, or it could trigger a strobe light in the area. “It can be flexible, depending on a community’s size and needs,” he says.
The various versions PBI has been testing have different ranges and different costs; generally speaking, the longer the range, the more expensive the equipment. But, asks York rhetorically, “do we need a super expensive and super long-range system that can spot objects 15 kilometers away? Probably not. It’s more important that I’m alerted if there’s a bear 200 meters to a kilometer away. That’s more actionable.”
So far, all the systems have proven highly successful at detecting objects. But the fine tuning—distinguishing between a polar bear and a truck, or even a fox or a dog—has been more challenging.
Part of the problem, York explains, is that the bears in Churchill, where PBI has been testing, are waiting for sea ice to form on Hudson Bay and so tend not to be highly active.
“They might flop down for 10 minutes, or they might flop down for six days,” he says. Once they go still, the radar drops their signal. “And the other side of it is polar bears are just uniquely soft, with their dense fur. So even though we think of them as being quite large, they're a large target with a soft kind of radar-proof cover.”
The solution to that problem is to teach the software to react to the right objects and ignore the wrong ones. Kirschhoffer explains that, during the COVID pandemic, he was able to sit at home in Bozeman, Montana, and react to signals from the radar outside Churchill.
Every time the radar saw something move a red light would turn on, and Kirschhoffer would log in to the server, look at the map, and look at the livestream. “And if it was a caribou, then I would mark it as a caribou. If it was a polar bear, I’d pin it as a polar bear.” With enough targets gathered, the AI should start to learn and try to apply that knowledge to new targets.
A multi-pronged approach
Nobody is arguing that bear-dar is the only solution to potential human-polar bear conflicts; it could, however, be part of a suite of tools that includes education, the development of alternative forms of waste management in Arctic communities—and, of course, a reversal of the global warming that is melting sea ice, exacerbating the problem in the first place.
“I think we as conservation organizations have a responsibility—especially when we're talking about conserving a large mammal that's a predator—to the people who live amongst them,” says York. “And so, that's one of the reasons we really are focusing more and more on coexistence and trying to make sure that communities have the resources that enable them to feel safer.”
None of the systems is cheap—the BYU system, which is the smallest, is designed to cost around $1,000 per unit. But Kirschhoffer is optimistic that once the AI is fully trained, attracting funding for deployment in vulnerable communities should not prove insurmountable.
“People want to build things,” he says. “People want to give money to something that is going to make a difference. And how cool if you had a tool that is truly effective and can make a difference and can save both bear lives and human lives?”
