The image streams over YouTube in crisp grayscale: a young cahow — known outside Bermuda as the Bermuda petrel — scrambles through a sandy tunnel and pokes its tiny head above the ground for the first time. It’s a few months old, but it has never seen daylight. Gray fluffball hatchlings spend their whole lives up to this moment in a pitch-dark burrow as far as 15 feet underground. Now, in the middle of the night, this little bird flaps and flexes its wings, perches at the edge of a cliff, and launches itself into the wind. It won’t touch down on land again anytime soon: a cahow’s first flight can last three to five years. While it may rest on the water for a few minutes here and there, it’s almost entirely airborne, zigzagging for hundreds of thousands of miles across the Atlantic high seas, even sleeping while in flight. If it survives this odyssey, it will come right back here — to this little speck of an outer island — landing as little as a yard away from the nest where it was born.
For centuries, no one knew this highly unusual bird still existed. The cahow, Bermuda’s national bird, was assumed extinct for centuries, and even after its rediscovery in the 1950s, its nocturnal life was a relative mystery. That is, until a Bermudian conservationist with a proclivity for DIY electronics decided to hack a couple GoPros and set up one of the earliest 24/7 livestreamed bird cams. There, he captured the unseen life of this critically endangered “Lazarus” species — one of the rarest on Earth — for the first time.
Today, Bermuda’s Nonsuch Island is the heart of the world’s only cahow breeding ground, a protected 15-acre home base to upward of 186 pairs. Jeremy Madeiros, warden of Bermuda’s Department of Environment and Natural Resources, does the hands-on work at this government nature reserve — monitoring nests, banding hatchlings, conducting health checks, and tabulating the data. And if you’ve ever seen him speaking about how it’s going, it’s likely that Nonsuch Expeditions founder and filmmaker Jean-Pierre Rouja, the GoPro hacker, is behind the camera. For the past 20 years, Rouja has been helping document the life and conservation of the Bermuda petrel — and piloting lightweight, scalable conservation tech in the process, using Nonsuch as his field station.
Like many Bermudians, Rouja grew up following the rewilding work on Nonsuch. In 2005, he set out to make a documentary short about the cahows’ comeback. His one big problem: the shooting circumstances were far from ideal. “They’re in dark, man-made boroughs,” he says of the birds. “You can access them, but then you’re ripping the roof off their house, and you’re not witnessing any natural behavior. We couldn’t afford to film underground. It just technically wasn’t possible at the time.”
By 2010, Rouja was done waiting around to see if someone would invent the gear he needed. “Basically, I ended up teaching myself,” he says of how he cobbled together the camera systems he needed — modular, waterproof, operable off the grid, able to auto-activate unobtrusively in the pitch dark, with lighting that would be invisible to the birds, and not to mention available at a grassroots conservation pricepoint.
A self-described “frustrated electrical engineer” with no schooling in electronics, Rouja knew there would be trial and error. So he joined the “whole subculture” he found online of people hacking the relatively new GoPro Hero, which he says were “the only cameras I could afford to risk destroying.”
On a work table strewn with electrical tape, jeweler’s screwdrivers, wirecutters, and a hot glue gun, Rouja cobbled together the custom camera system of his dreams. He started by removing two GoPros’ IR filters, so they would be able to pick up infrared underground. Then, he built his own light arrays with individual military-grade, 940-nanometer micro-LED bulbs, rigged with custom, laser-cut faceplates and transformers that would enable them to run off any power source (a car battery, for example) in the wilds of Nonsuch. (He says wrote to GoPro, hoping for some kind of sponsorship or accolades for his creativity. To his surprise, the company wasn’t at all pleased to hear he’d figured out how to turn its cameras into stealth spying devices — even if the subject of his surveillance was an endangered seabird. (GoPro has not responded to a request for comment at the time of this story’s publication.)
By 2011, Rouja and Madeiros had launched one of the world’s first 24/7 livestreamed wildlife cameras, raising awareness and delighting bird nerds by sliding them neatly into the standard four-inch PVC pipes they’d built into the birds’ subterranean burrows, capturing the scene from above.
With most wildlife cams, Rouja says, “light is blasting at the bird, and the bird is looking in a box.” His cameras were designed to allow viewers not to feel like unwanted intruders. “We get very natural behavior,” he says. And despite the darkness, you can see it all vividly in black and white, down to the details in the feathers.
A several-year partnership with the Cornell Lab of Ornithology helped cement the livestream as a worldwide birders’ favorite. At last count, Rouja says, more than 40 million minutes of cahow video had been watched.
The cahow cams proved fun for spectators (and alluring for donors), but they’ve also driven the research forward. These birds are “elusive and difficult to study,” as Madeiros puts it. Filming them around the clock allows the team to “ground-truth” preexisting theories about breeding and behavior, Rouja says, and discover new, wild truths in the process.
Madeiros boats over to Nonsuch Island every few days. With the cahow cam, the team can see everything that happens in between, from the odd but harmless storm petrel interloper (cahow cam watchers named him “Stormy”) that repeatedly made its way into burrows attempting (and failing) to woo and canoodle with cahow chicks, to the revelation of a symbiotic relationship between the cahow and one of the world’s rarest lizards, the Bermuda skink. (They cozy up with the cahows to keep warm in the winter and reciprocate by scavenging to keep the burrows tidy.) Last fall, the team even caught a cahow in an act of infidelity on camera — and all the drama that ensued.
Rethinking the cost of conservation tech
To Rouja, the work on Nonsuch isn’t just about saving one rare bird species. He’s beta-testing conservation fieldwork tech that could be put into play anywhere.
Most field tech — satellite trackers, thermal cameras, deep-sea sensors — was originally built for the military, oil exploration, or commercial science. A single device can cost upward of $20,000. For most conservation projects, that kind of price tag is a liability, he says. “You need gear you can afford to lose” because “a lot of stuff you put out [in marine research] just doesn’t come back.” Furthermore, he says, that level of sophistication is often overkill. “The fact is that you can probably build it for 300 bucks, or 500 bucks, depending on what it is.”
“We’re using Bermuda as a proof of concept to make sure these technologies work, with the goal of then being able to roll this all out at scale.”
The gen-one cahow cams were proof of concept. Rouja is encouraging conservationists elsewhere to follow his blueprints for observation of other underground species. Hawaii is likely the next roll-out location for his nest cams. And as the cofounder of blue tech rapid development facility Station B, he’s working with partners like Woods Hole Oceanographic Institute, Cornell, and MIT, testing and hardening gear for coral reef and ocean sensors, marine acoustics, and soundscape monitoring that he hopes will be affordable enough to deploy en masse.
For example, if rats made it to Nonsuch and took hold during nesting season, they could easily wipe out an entire generation of cahows and throw the fragile species off its track to recovery. To guard against this in the past, the team relied on volunteers watching motion-detecting trail cam livestreams all day and night, in shifts. But now, they’re working with the Nature Conservancy to trial an AI-powered rodent detection system. For the past year and a half, human volunteers have been helping train the Nature Conservancy’s machine-learning platform to differentiate between footage of harmless animals and threatening ones. Soon, Rouja hopes, these cameras will automate the search, so they can spot a rat before its presence would become a “disaster.”
“The cahow project was my gateway into conservation tech with it all now running in parallel,” he says. “We’re using Bermuda as a proof of concept to make sure these technologies work, with the goal of then being able to roll this all out at scale.”
In the meantime, they’re up and running on Nonsuch, where multiple cahow cams will be streaming and recording this cahow hatching season in 4K HD, continuing to broadcast a species’ journey back from the brink.
