02.04.2021 08:00 AM

How to Make a Fitbit for an Elephant

The accelerometers give scientists information about whether animals are swimming, walking, running, or even sprinting up a hill.

elephant walking — Photograph: Getty Images

When Daniella Chusyd was in graduate school at the University of Alabama, Birmingham, she noticed that many of her colleagues used step counters like Fitbits to study obesity and activity among people. She wondered if she could use the same method in her own research, which examines how obesity and metabolism affect reproductive health. There was only one difference: Chusyd studies elephants.

Unlike researchers who focus on humans, Chusyd couldn’t use consumer fitness trackers that were ready to go right out of the box. For one thing, elephants are giant. So she had to design a mounting system that would fit the step counters, also called accelerometers, firmly but comfortably around an elephant’s enormous legs. “They are pretty massive. What fits around their wrist or ankle fits around my waist,” she says. Chusyd designed large adjustable bracelets and used lots of zip ties to secure the devices, which were also placed inside waterproof boxes and wrapped in several industrial-strength plastic bags to protect them from elephantine bathing habits. After everything was waterproofed, wrapped, and secured, the bracelets weighed about six or seven pounds each.

For the most part, the zoo animals were game, though Chusyd lost one device after an elephant used its trunk to rip it off and step on it. Chusyd, now a postdoctoral fellow in the School of Public Health at Indiana University, says that because this was such a novel use of the devices, she also had to validate all the data by watching the elephants as they walked, counting their steps, measuring their stride lengths, and matching those observations up with the accelerometer data. It took hours of work to make sure the sensors were accurate and to figure out the right signatures in the data that pinpointed when the animals were walking or digging. But Chusyd says these sensors produced incredible data that researchers wouldn’t otherwise be able to get. “You gain insight into animal behavior that you wouldn’t be privy to otherwise, because it’s impossible to follow them 24 hours a day, seven days a week, or see them all the time because of their environment,” she says.

While Chusyd and other scientists don’t use commercial counters like the Apple Watch or the Fitbit, the technology is essentially the same. Step counters are triaxial accelerometers, which use electromagnetic sensors to pick up movement across three planes. If a wearer’s foot moves up and down, side to side, or backwards and forwards, the device senses that movement and interprets it as a step. And over the last 20 years, many animal researchers have discovered that accelerometers can be used for far more than notching up step counts. The devices have been used to study penguins, cormorants, badgers, pumas, and polar bears.

“The thing about accelerometry is it allows you to look at animal behavior,” says Rory Wilson, a professor of aquatic biology at Swansea University. He says these sensors can help provide information about whether animals are swimming, walking, running, or even sprinting up a hill, and they give scientists important proxy measurements about how much energy animals are using up during these activities. “For someone who works with animals in the wild—and half the time you can’t see them—you have a really powerful tool, amazingly powerful tool,” he says.

Wilson first heard about using accelerometers in 1999 when Ken Yoda, a professor of behavior and evolution at Nagoya University in Japan, published a paper in the Journal of Experimental Biology about using the devices to track swimming penguins. Wilson was skeptical. “I sort of remember saying to someone, ‘Yeah, that’s quite cute, but I don’t see how that would be useful,’” he says. But he slowly came around. He started using the sensors to study the angle at which seals dive into the water. Then he started to get curious about using the devices to gather data that could infer animals’ energy expenditure. “Energy expenditure is the ‘money’ for an animal,” he says. “How much money have you got? How much energy have you got? How hard do you have to work to get energy back?”

Understanding energy expenditure can help scientists understand how well animals are doing and whether they are going to be able to hunt, reproduce, and survive. By now, Wilson has used accelerometers to study all sorts of animals including sea turtles, sheep, bats, hawks, and penguins. He combines the accelerometer data with inputs from other sensors that measure temperature, magnetic force, and geolocation to understand exactly what the animal is doing and where it is. For example, the technology allows him to track penguins as they sit on their nests, get up, waddle to the ocean, and dive in. His sensors can stay on the animals for weeks, and after he retrieves the devices, Wilson can follow along as the penguins swim and dive and fish, all from thousands of miles away.

He’s gotten so good at reading the data, he can even start to understand details of the animals’ physical state. He can tell when the penguins are full of fish, for example, because that changes how they waddle. Or he can tell when a horse is starting to walk over tricky terrain. “That’s really cool stuff,” he says.

Like Chusyd, Wilson has become well-versed in figuring out how to attach accelerometers to animals and making sure the sensors will survive the data collection process. For penguins and other birds, he’ll insert special tape under their back feathers, creating a little waterproof pocket into which he encloses the device. He used magnetic and spring-based clips to attach sensors to sharks’ fins. When he studied sheep urination, he cut little holes in the wool on the animals’ rear ends, glued the sensors into their coats, and repacked the pockets with the tufts of shaved wool. For sloths, he used a harness, and for bats he used rubber cement to affix the accelerometers to their leathery skin.

For Anthony Pagano, a postdoctoral researcher who works with the US Geological Survey, accelerometers have helped illuminate the activity of polar bears living north of Alaska, providing insights that can be nearly impossible for humans to observe. “We have a lot of detailed information about changes in body mass and survival rates, but we don’t have very much information about basic movement patterns and what their basic behaviors are on the sea ice,” he says.

These bears live in extreme and remote environments. Temperatures can shift from 40 degrees below zero up to 30 degrees above zero. The polar bears are diving in and out of frigid salt water oceans, hanging out on ice floes, and tramping around on solid ground, too. Pagano ultimately had to encase the accelerometers in epoxy to make them waterproof, mounting them in an aluminum housing and bolting the whole unit to tracking collars around the bears’ necks. Like Chusyd, he also had to figure out what the patterns in the data meant by fitting bears in captivity, observing them, and then matching up those observations with the data from wild animals. Between figuring out the right bolting system and validating the data, it took Pagano a year to get ready to put the devices on bears in Alaska.

Accelerometers have some limitations. Because Pagano relies on collars to attach his sensors, he can only tag female bears; males have necks that are larger than their heads, meaning the collars will slip off. And where the sensors are placed on the animal is really important, especially if scientists want to study a particular motion or behavior. At first, Pagano wanted to find movement patterns that could identify when the bears were killing and eating seals. But because he had to attach the accelerometers to neck collars, the sensors haven’t been able to find distinct signature patterns for those killing and eating movements, because those motions are happening in other parts of the body, like the hands and feet. Maybe if the accelerometers were attached to the bears’ paws, they would be able to find that data—but there are just too many other head movements that the animals make for the sensors to pick up hunting-and-eating-specific signals.

Wilson also points out that accelerometers lose a certain environmental perspective because they can’t tell you if there are other animals around the one you are observing. It might be useful to know if a particular or unusual motion is happening because the animal is standing in a crowd or watching another animal approach.

These devices are also limited by how much data they can store and by battery life. Accelerometers don’t relay information in real time, like GPS does, so researchers have to collect them after six weeks or so in order to gather the data and recharge the devices.

Still, Pagano says the sensors have been great for identifying when polar bears are swimming, walking, or resting, and how much energy they’re using to do all those activities: important information that will help researchers start to make predictions about how well the bears might do in the future as their habitats change. And the devices allow humans to observe animals without affecting their environments in ways that could potentially influence that animals’ behavior.

Observing animals by eye is traditionally the gold standard for researchers, says Danielle Brown, a lecturer in the biology department at Middle Tennessee State University who has used accelerometers to study anteaters. “But we have this problem: That as soon as we’re there, we’re changing their behavior,” she says. A human presence might discourage predators from approaching or make the animals hesitant to relax. “There’s a lot of behaviors that are very context dependent,” says Brown. “When humans are there, that changes the context.”

And she says that while accelerometers may miss some behaviors, they may also pick up on movements the human eye doesn’t catch, either because they are subtle, or rare, or happen in the dark.

As technological advances make accelerometers and their batteries smaller and longer lasting, and as scientists start to pair them with other sensors, Wilson believes the devices will become more and more pervasive in research. “I think the most important thing to understand about accelerometers is that we haven’t begun to scratch what we can do with them,” he says.