On the Trail of Sea Life with an Autonomous Underwater Vehicle

EarthzineAutonomous Underwater Vehicles 2017, Original, Themed Articles

A look at SharkCam and TurtleCam, designed to help us bring threats to marine life into focus.

Marine biologist Kara Dodger (left) and WHOI engineer Amy Kukulya partnered to develop and test SharkCam autonomous tracking and imaging system to follow and film leatherback turtles. Image Credit: SharkCam, WHOI

As a researcher and engineer at the Woods Hole Oceanographic Institution (WHOI), I’ve operated autonomous underwater vehicles (AUVs) in some surprising and difficult places. So, when Tom Austin, the head of the Oceanographic Systems Lab (OSL) at the time, came to me and said, “Amy, I’ve got just the project for you,” I knew it had to be interesting.

It turned out to be more than I could have imagined. The Discovery Channel was asking if we could refit one of our REMUS-100 AUVs into something that would eventually become known as SharkCam and could track and film live animals like great white sharks in the wild for a program to air during its annual Shark Week.

It was clearly going to be a challenge, and I wasn’t alone in my doubts that we could pull it off. Greg Skomal, who heads the Massachusetts Shark Research Program and was eager to learn more about sharks in the wild, wasn’t convinced we could do it, either. Even Nick Stringer of Big Wave Productions, the would-be director and head cheerleader of the show in meetings with the Discovery Channel, knew it was a gamble. To pull it off, we would need to modify a REMUS (an acronym for Remote Environmental Monitoring UnitS) to do a job beyond anything it had ever been called to do, and I’ve made them swim under ice, patrol the edge of a glacier, and hunt for the lost eighth wonder of the world.

The payload nose of SharkCam-600, a REMUS-600 meter AUV configured to follow a tagged shark, with omni-directional navigation, LED lights and cameras. Image Credit: Amy Kukulya, WHOI

A Long Time in the Making

The concept of autonomous tracking of a live marine animal began in 2008, when WHOI biologists approached OSL and inquired about REMUS’s capabilities. How amazing would it be, they asked, if a REMUS could follow a basking shark to find out if the planktivorous fish actively sought out plankton “clouds”—areas of high plankton concentration.

With a small WHOI grant, REMUS engineer Gwyneth Packard engaged lead software developer Roger Stokey to develop a navigation algorithm and simulate whether a vehicle could follow a moving target, in the form of a transponder emitting acoustic signals REMUS could home in on, and predict where the transponder might be in the future and act accordingly.

Straightforward as it might seem, it’s a complicated problem. The vehicle would have to constantly re-calculate its own position in the water, as well as the speed, depth, and direction of the moving transponder in order to place itself immediately behind, above, below, or to the side of its target.

After tests in the murky waters off Cape Cod, first following Skomal standing in for a live animal on an underwater “scooter” and then tracking four sharks through the noisy inshore waters of Chatham, Massachusetts, SharkCam was born. We then took it to Guadalupe Island off the Pacific coast of Mexico in 2013 for its first deep-water science expedition. The clear waters around the island are famous for its seal colonies—and the many white sharks that come to feed on them. Skomal and our new colleague, Mauricio Hoyos Padilla of Pelagios-Kakunja A.C., a marine conservation organization based in Baja California, wanted to use SharkCam to learn anything they could about the natural behaviors of sharks below the surface. Up until that point, subsurface observations had not been possible.

We tagged a shark, launched the vehicle, and then huddled around the computer receiving occasional updates from the vehicle. We watched, mystified, as it cruised along normally, and then suddenly changed altitude, attitude, and orientation several times. A leak alarm signaled the premature end of the mission and the vehicle slowly floated to the surface.

As we hoisted SharkCam on board, the cause of its acrobatics, and the leak, suddenly became apparent: Deep gouges in the paint showed that one or more sharks had attacked the vehicle in the water.

In all, we collected 13 hours of video tracking four sharks. The most surprising thing to arise was the fact that sharks routinely dove well beyond the vehicle’s 100-meter limit, something that Skomal and Hoyos hadn’t expected. It even appeared that the sharks were intentionally diving deep, hunting on the edge of darkness, so that they could silhouette their prey (REMUS or something edible) against the sunlit surface waters and then attack from below. It turns out REMUS looks a lot like a yellowfin tuna.

Footage of sharks attacking the vehicle quickly became the hit of Shark Week and went viral on the internet. But the SharkCam team had our sights set on what the sharks were doing in the deep water, and for that we’d need new tools.

In 2013, a team from the Woods Hole Oceanographic Institution took a specially equipped REMUS “SharkCam” underwater vehicle to Guadalupe Island in Mexico to film great white sharks in the wild. Video Credit: WHOI

The following year, we went back to Guadalupe, this time with two REMUS vehicles, one rated to dive to 600 meters and outfitted with lights. Our goal was to find out what the sharks were doing below 100 meters—and also to reduce attacks on the vehicle.

We were only partially successful.

While we were able to get down to the level of the deepest shark dives and even film what Hoyos thinks may be the first recorded instance of a shark swimming while “asleep” (white sharks don’t really ever sleep in the traditional sense, but they do glide to rest and conserve energy), the vehicle continued to prove too tempting a target for hungry sharks, and we recorded nearly a dozen attacks at almost every depth, as well as several different hunting techniques.

Some sharks even displayed territorial behavior toward the vehicle, indicating that, in addition to prey, they also considered it competition. All of this demonstrated, to shark expert, engineer and filmmaker alike, that the more we open the ocean to observation, the more we’ll learn—and the more questions we will inevitably ask.

A New Target

When biologist Kara Dodge approached me in 2015 to ask if I could modify SharkCam to follow a sea turtle, it was as if we were faced with an entirely new ocean. Unlike sharks, which tend to follow a relatively straight path through the water, turtles yo-yo as they alternately dive to feed on jellyfish and then surface to breathe. As a result, we had to develop entirely new navigation algorithms. We also needed to modify the tag with suction cups to stick on (and release from) a turtle’s shell.

A tagged loggerhead sea turtle surfaces to face one of the many hazards they find in Martha’s Vineyard Sound: boat traffic. Image Credit: TurtleCam, WHOI

In September 2016, we succeeded in tagging our first leatherback sea turtle and following it for more than six hours, and TurtleCam was born. That trip and eight that followed resulted in a trove of video that Dodge is still combing for insights that will help shed light on things such as the respiratory and metabolic rates of turtles, what and how much they eat, and the nature and diversity of hazards that endangered sea turtles encounter in the waters around Cape Cod.

Dodge also has information from a second vehicle we deployed to gather temperature, salinity, biological productivity, and bathymetric data in the water near the tagged animals. This will help her place the behavioral data into the context of the animal’s habitat.

Although many North Atlantic populations of sea turtles are stable, their primary threats—entanglement, boat strikes, plastics, coastal development, and climate change—are human-caused and so are not likely to diminish any time soon. We only have to look to parts of the Pacific, where almost all sea turtle species are in trouble, to see how dire things could become.

It’s the nature of any problem, however, that the solutions don’t appear unless we look, and that is exactly what SharkCam and TurtleCam are designed to do—help us look beneath the surface into the world of marine life that is so alien to us on the surface. These robots and the video footage they return are slowly bringing that world into focus and making it, and the problems that sharks and turtles face, more real to scientists and the public than it has been in the past. Answering the questions that inevitably arise as we peer longer and deeper into that world will undoubtedly take time, but with the help of robots like these, that world is a little less alien and the animals that share the Earth with us are a little more familiar.

Amy Kukulya is an ocean vehicle operations engineer and principal investigator at the Woods Hole Oceanographic Institution (WHOI). Her research interests are navigation, acoustic communications, underwater archaeology, three-dimensional tracking of marine animals and under-ice capabilities.

Ken Kostel is a science writer at the Woods Hole Oceanographic Institution who also photographs, produces video, does social media, blogs from the field, and does just about anything else to get the word out about ocean science and exploration and the interesting people who do the work.