At first glance, Andrew Wilson’s latest research seems risky. “I’ve been counting birds for a living for a long time,” he says. So why explore whether drones equipped with audio recorders can do just as well as his own ears? Wilson isn't worried about being replaced by machines—but he does think the assemblies could help ornithologists research songbird populations, especially on terrain that’s tricky for scientists to navigate, such as wetlands.
Ecologists have enlisted drones on a whole range of projects, from counting species in photographs to collecting radio-collar data from individuals and helping plant trees. But the work by Wilson, an ecologist at Gettysburg College, is the first ecology project using drones to capture audio data. The results were published last week in The Auk. “This is kind of baby’s first steps, but we think we’ve proved concept,” Wilson says. He’s particularly pleased that buying all the equipment only took about $1500, so it won’t break a research team’s budget.
For the trial run, Wilson stuck to familiar territory. The team used the campus atheletic field, and placed speakers emitting birdsong recordings at measured intervals. Then they flew the drones by the speakers at different heights to test how close the recorders need to be to pick up useable data. This also gave the team a chance to figure out how best to hang the recorder from the drone. Dangling the recorder about 26 feet, they found, reduced enough noise from the motor while also avoiding any potential tangles with obstacles below.
Once Wilson knew the recorders’ range, he pitted the drones against a trained ornithologist skilled in counting birds by ear—himself. Although some researchers are already training computers to identify birds based on recorded calls, that technology is still too limited in reliability for field surveys. So for this work, humans still have to listen to all the recordings after the flight to gauge the audio equipment's success.
In the head-to-head comparison on a wide grassy area studded with trees, Wilson’s ears on the ground picked up more songs than the drones, reassuring that field surveys aren’t going away any time soon. “I don’t think it’s ever going to replace what field ornithologists do,” Wilson says. “I think it’s going to be an additional tool.”
But for a first attempt, he was pleased with results from the flying recorders, especially since all bird-counting protocols—especially those aimed at small birds—have their weaknesses. In three minutes at each location, he identified birds of 51 species; the drones that followed him allowed the team to identify birds of 32 of those species. The cheaper, lighter recorders they chose for practicality’s sake were likely responsible for some of the missed birds, about half of which Wilson himself only heard one individual.
The real perk of drones would be saving scientists the trouble of getting out to a specific point on the ground. “One of the big problems with bird surveys is walking from one point to another,” Wilson says. But with a drone, he adds, “you can really zip between them and potentially cover a large area.”
Wilson did stumble on a couple of hiccups unique to such a study. Gray Catbirds are so common in the area that experts listening to the recordings had trouble pinning down precisely how many birds were calling at once. And Mourning Doves, it turns out, call at about the same pitch as the whirr of a drone motor, so they tend to get drowned out.
This is one challenge that could be reduced with advances in recording and drone technology—or even a different type of vehicle. Despite dangling 26 feet below the drones, the recorders still picked up some of the noise from the motors that drive the quadcopters. “It’s a very classic little drone, but it makes a lot of noise,” says Julie Linchant, an ecologist who has reviewed projects enlisting drones for conservation and science goals. Linchant suggests a balloon could be particularly helpful for audio recording, since it can hover quietly.
Wilson is the first to say the technique needs more work. He’s particularly concerned with drones might affect how birds behave. “If the noise of the drone is too loud, then birds aren’t going to waste the energy singing,” he says. His goal for this summer is to make audio recordings on the ground starting before a drone flies by, to see whether birds hush up when drones are near. After all, no matter how advanced the drone or audio equipment, his technique can’t record silent birds.