Tech

How a snot-collecting marine drone is giving us an exciting new view of whale life

Jul 10, 2018 /

A flying device combined with some nifty software is serving up invaluable information about the health of whales — and our oceans.

Back in 2011, researcher Iain Kerr was having a frustrating day chasing sperm whales on the Gulf of Mexico. Kerr, CEO of the nonprofit Ocean Alliance and chief scientist of Parley SnotBot, would wait to spot a whale surfacing. As soon as he did, he’d tell the crew to bring the research vessel to the mammal as quickly as possible so a researcher could shoot it with biopsy darts and extract a plug of skin and blubber to be analyzed.

That day, the researchers found themselves rushing after surfacing sperm whales only to watch them dive before they could reach them — the whales can plunge as deep as 3,800 feet and stay below for more than 90 minutes. And this kept happening.

As Kerr sat in the bow of the boat and contemplated the day’s failure, a sudden cloud of whale blow — the spray exhaled by a whale — settled over him. And he had an unpleasant but scientifically fruitful realization. “It’s pretty fishy, pretty stinky stuff,” he says. Kerr knew the stink meant that biological material, not just water and air, was present. Which meant it might be worth sampling. But how could he get close enough to surfacing whales, and quickly enough, to catch the blow before they dove? Kerr, a drone aficionado, had an idea: why chase down a 4–6-mile-per-hour whale that can dive at any moment by using a boat that goes 8 miles an hour, when you could launch a drone that flies at 50 miles an hour?

@christian_miller_photo; Courtesy of Parley for the Oceans/Ocean Alliance

Kerr began working with students at the Olin College of Engineering to design what they ended up calling the Parley SnotBot: a drone that could be launched from a research vessel, hover over a whale, and collect its blow. Although the technical term for the blow is “exhaled breath condensate,” Kerr calls the substance “snot” since it also includes the lubricants in a whale’s lungs and nasal tracts, along with other cells that get blown up and out. After trying to design their own drones and experimenting with the best ways to collect snot — including sponges, petri dishes hanging from poles and veil-like cloths — the team settled on a commercial drone mounted with petri dishes. (The drone was funded at first by a Kickstarter campaign and later in part by the collaboration network Parley for the Oceans.)

Since 2015, the Ocean Alliance has deployed SnotBot on seven expeditions. It’s collected over 500 samples of five species of whales in three areas — southern right whales off the coast of Patagonia; humpback whales and orcas off Alaska; blue whales in the Sea of Cortez in Mexico; and gray whales on the Pacific Ocean side of Baja California. On the expeditions, a researcher on the boat pilots the drone into the blow of a surfacing whale. While collecting snot, the drones also transmits photo and video of the whale back to scientists; when the whale dives, the pilot brings the drone home and the snot is harvested from its petri dishes. The team has also used drones to collect whale poop, drop underwater microphones, and fly over protected waters with a night-vision thermal camera to look for poachers.

The drone’s ability to take hi-res photos of the whale has been critically important and informative. A whale’s fluke, or its tail fin, is the species equivalent of human fingerprints — its particular color, scars and serration are unique to that animal. Previously, researchers took photos of whales’ flukes and then reviewed them one by one to match them to known subjects. But with this drone, that’s all changed. Special software designed by scientists at Intel harnesses artificial intelligence that instantly analyzes a fluke shape, compares it to images amassed from decades of whale research by the Alaska Whale Foundation and Happywhale, generates an ID, and brings up any existing data on the animal.

The software also gives an instant health check for the whale, even if its body is partially submerged. When it analyzes the photo, it’s able to estimate the mammal’s 2-D shape and provide a score known as Body Condition Index for researchers. This score can tell researchers on the spot if a whale is healthy or malnourished. Again, this immediacy is invaluable. “By arming researchers with this advanced ability to immediately understand and assess the state of things when they’re out in the field, they can take in the rest of the environment and put everything in context,” says Ted Willke, Intel’s senior engineer on the project. “Before, they’d have to go back to the lab and then realize after looking at photos that there was something else they should have been paying attention to.” He says the next step will be to train the software to recognize pregnant whales and juveniles and to identify where exactly on the whales they’re carrying fat, which will give researchers a greater understanding of how and when whales lose and gain blubber over the course of their feeding cycle.

@christian_miller_photo; Courtesy of Parley for the Oceans/Ocean Alliance

Thanks to the drone, researchers can gather many more samples from whales — without disturbing them. University of Alaska–Fairbanks marine scientist Shannon Atkinson specializes in studying hormones in whales. She depends on getting snot and fecal samples to analyze. She echoes Kerr’s frustrations with traditional whale research: “You’re sitting around waiting for a whale to poop,” she laughs. “If they don’t have to go, you don’t get anything.” Before she had access to the drone, she collected about 110 blubber biopsies and 45 fecal samples in 10 years of expeditions. (Snot, even though it does contain hormones, was not collected before Kerr had his epiphany.) In a single 2017 expedition with the Parley SnotBot, Atkinson got 39 snot samples, and she’s collaborating with Ocean Alliance to analyze the hormones they contain. She won’t be retiring the pooper-scooper just yet; whale poop, rich in information about diet and nutrition, is useful in its own right. Another major advantage to the drone method is its noninvasive nature. Shooting biopsy darts at a whale usually causes it to dive down and hide — and, in some species, if one whale startles, that can scare the whole pod away, which means a researcher might get a sample from only one whale. So far, Kerr says there’s no evidence that the drone bothers the whales. “They’re quite used to having seagulls and birds flying around them,” he points out. And he believes the drone’s relatively high-frequency sounds don’t penetrate the water.

The hormones in whale snot will allow scientists to study whale pregnancies and stress levels. Atkinson’s lab has been able to isolate five different hormones from the blow: testosterone, progesterone, aldosterone, corticosterone and cortisol. Of special interest is progesterone, the hormone that allows mammals to maintain a pregnancy. Elevated amounts of it in blow indicate an expecting whale. (Fun fact: whales carry their young for 10 to 18 months before they give birth.) Atkinson hopes to track the reproductive rates of whale populations in order to monitor their overall health as a group. Examining levels of cortisol, a stress hormone, could show researchers how whales are reacting to, say, climate change or ocean noise.

@christian_miller_photo; Courtesy of Parley for the Oceans/Ocean Alliance

What else can be found in whale blow? DNA! Scientists are using the genetic information to count individual whales and even trace matrilineal lines. Oregon State University conservation geneticist Scott Baker says that although the amount of DNA in blow is relatively low (compared to, say, blubber), its quality is high; along with water, whales exhale epithelial cells from their lungs. That’s good news for monitoring populations as they recover — or don’t — from their decimation due to commercial whaling in the 19th and 20th centuries, which Baker calls “one of the great challenges for whales and dolphins around the world.” Getting greater numbers of DNA samples means being able to get accurate counts of whales in different parts of the ocean.

This drone-enhanced view into the lives of whales gives us a window into how our oceans are doing. “We often think of whales as sentinels of the health of their ecosystem, because everything in that ecosystem is affecting them,” says Atkinson. Her lab is just beginning to analyze the drone-collected samples; she wants to study how pregnancy rates may have changed in humpback whales due to warming events in the Gulf of Alaska and how stress hormones may factor into that. Using the image-analyzing software to measure the “fatness” of a whale and then connecting that with stress levels and whether it’s pregnant or not, year after year, will give scientists a rich trove of data on individual whales. Look at those data points for entire populations of whales, and you can get an idea of what’s going on in the ecosystem on a larger scale. “That’s the kind of information we’re going to need, to understand the changes in the ocean over the next decade,” says Baker.

The drone has led to exciting moments for researchers. On a 2017 expedition in the Chatham Strait area of southeast Alaska, Willke was with the team on the boat as they listened to whale calls via underwater microphone. Suddenly, Alaska Whale Foundation marine biologist Fred Sharpe thought he heard a call that sounded familiar. Could it be Trumpeter, a whale he hadn’t encountered in 23 years? A drone snapped a photo of the whale from the air, and the software ran an analysis of the image against an old photo of Trumpeter: it was a 92 percent match. “It was astounding,” Willke says. Before the drone and AI technology, “there was no way they could have matched that animal.”

From the drone’s-eye view, says Kerr, “it’s a whole different world.” He explains, “I’ve got friends who are looking at my pictures saying, ‘I never knew they did that!’ and these people have been studying them for twenty years.” Scientists have gotten an up-close look at how humpback whales catch prey. Or, in a sweeter example, the Parley SnotBot has captured a never-bef0re-documented shot of a mother humpback whale stroking its calf with its pectoral fin. This kind of intimate, up-close footage has the potential to break through the cloistered world of researchers and inspire more people to feel connected with whales. In an interview with Oceans Deeply, Kerr asks, “Is this changing the field of biology? No. But does it make it easier for us to empathize with this creature? I think it does.”

@christian_miller_photo; Courtesy of Parley for the Oceans/Ocean Alliance

Thanks to the drone, marine research is being revolutionized — at a bargain price. In summer 2018, the team plans to be off the coast of Massachusetts studying humpback whales, and later this year, they plan to go to Gabon, West Africa, to assess whale populations in an area where shipping traffic is increasing. It would be difficult and extremely expensive to get a fully staffed research vessel (which costs tens of thousands of dollars a day) to these waters, says Kerr, but a drone costs a few thousand dollars, a price low enough that the team will be able to leave one behind for people in Gabon to use. “Oceanography has, on many levels, been a prerogative of the privileged,” says Kerr. Drones can change that, opening up marine research to citizen scientists and to countries that can’t afford traditional programs. He adds, “I think we’re democratizing science with these drones.”