When Eddie Game straps the first wallet-sized recording device around a tree trunk in Papua New Guinea, he’s so deep in the Adelbert Mountains it’s a three-day hike from the closest road. A second device is secured on a neighboring tree. One will capture ultrasonic sounds imperceptible to human ears, the other, audible sounds. Local rangers swipe bush knives through the underbrush, guiding Game and his team of conservation scientists through this rugged, largely untouched forest. Their shoes remain perpetually soggy. They’ll arrange devices in four more areas today and set them all to record for 24 hours. Hopefully, rats and possums won’t chew up the microphones or straps.
“What we’re interested in is overall soundscapes, how rich they are,” says Game, lead scientist for The Nature Conservancy’s Asia Pacific region and senior fellow at the University of Queensland, Australia. The more complex and dense the sounds, the healthier the habitat. Game is in Papua New Guinea to measure whether or not the conservation zones The Nature Conservancy helped set up 15 years ago are truly benefiting wildlife.
Bioacoustic technology has long been used to identify and count specific species. Game and his team, however, use holistic soundscape recordings to monitor overall animal biodiversity in tropical forests. They recently published the results of their work in the journal Conservation Biology, discovering that the dawn and dusk choruses were the most impacted by human activity. In Science, the team stressed the incredible potential of this research method, noting their random forest sound models “can predict species richness with very high accuracy.”
Though tropical rainforests cover a mere 2% of the Earth’s surface, they are home to more than half of all plant and animal species. During peak noise-making hours, these forests practically vibrate with sound: Birds call, insects whirr, monkeys howl, frogs chirp, wildcats growl, bats click, rodents squeak, and snakes hiss. The soundscape changes dramatically, however, when biodiversity decreases.
“There’s a very clear signal when an environment gets degraded — it gets quieter,” says Game. “Typically, there’s a big peak in sound at dawn. But the more a forest is degraded, the lower that peak is. We’ve been to half a dozen countries and we see the same thing: the loss of the dawn chorus. It’s so dramatic to us to hear it.”
Using sound to map a forest’s biodiversity may overcome many of the pitfalls of conventional measuring tools. Satellite imagery only tells us about the forest canopy. And in selectively logged forests, the canopy might not even change. Inventory surveys and trapping programs are immensely time-consuming. Wildlife cameras only capture a tiny slice of the forest’s inhabitants. Sound can fill in many of these gaps.
“Acoustics are better than some other methods because you are getting a robust signal and it doesn’t matter who the observer is: I could place a microphone or you could place a microphone, and we’d both capture the same thing,” Game says.
The Rainforest Connection, a San Francisco-based nonprofit organization, uses bioacoustic technology specifically to listen for sounds of illegal poaching and logging in Sumatra, Cameroon, Ecuador, Brazil, and Peru. If its network of pre-owned smartphones hears chainsaws, gunshots, or trucks in protected areas, it alerts local rangers. Bioacoustics are also used to identify both endangered and invasive species.
The same technology is being used in coniferous forests in the U.K. and in oceans the world over. By making repeated recordings, the Acoustic Ecology Lab at Arizona State University is capturing the effects of climate change on the sonic landscapes of the American southwest. Lab co-director Garth Paine notes that sound is a critical environmental signifier, and changes can often be heard before they can be seen. Any recordings made in undisturbed forest areas can serve as a baseline for future studies of the human impact on biodiversity.
Back at camp in Papua New Guinea, Game sits barefoot on the hut floor wearing headphones. As he listens to the recordings, he is blown away by how much more complex it sounds than when he was out placing the equipment.
“Ears can be tricked,” Game says. Our bodies are adept at filtering out sounds outside the typical frequencies, deeming them background noise.
The sound data begins to dance across his laptop screen; on the false-color spectrogram, each species’ sounds are visualized as brightly colored waves. Different species occupy different frequencies—likely so they can better communicate with their own kind. In a healthy forest, noise will be spread across the entire sound spectrum.
“It’s so dramatic to us to hear it.”
Game’s research colleague Zuzana Buřivalová is an assistant professor at the University of Wisconsin and principal investigator of the newly formed Sound Forest Lab, which uses sound recordings to monitor biodiversity and conservation effectiveness in tropical forests. She likes to imagine rainforests as an orchestra in which each species is a different musical instrument. When she listens, she hopes to hear a lush, vibrant, and varied acoustic composition.
“If you think of an orchestra, you have instruments that play at different frequencies. If, let’s say, the violinists didn’t show up, we would see a gap in the soundscape. For me, those instruments are different species, and I can calculate how saturated the soundscape is at any moment,” Buřivalová said last year.
Intensive agriculture and logging can decimate habitats and destroy delicate ecosystems. Large swaths of Amazonian forest have been lost to cattle ranches and soybean farms. The diverse forests of Southeast Asia cleared for palm oil plantations. Many logging operations adhere to the guidelines of forest certification programs that aim to balance economic viability and environmental sustainability, but there’s no standard way to compare or measure the effectiveness of different programs’ strategies.
Conventional measurement methods were showing that the biodiversity of logged forests was similar to unlogged forests. Skeptical of the sensitivity of these methods, Game turned to bioacoustics. To compare logged and unlogged forest diversity in Borneo, his team arranged a grid of microphones approximately 1 kilometer apart in both types of forest. When defining an area’s biodiversity, scientists look at both alpha diversity, which is the number of species, and beta diversity, which is how those species communities differ over land.
Game discovered a troubling trend: homogenization. “In unlogged forests, every kilometer of audio sounds completely different. In logged forests, it all sounds kind of the same,” Game says.
Sound recording devices are increasingly inexpensive, but the tech that translates that information into usable data is pricey. Fortunately for researchers like Game, this isn’t a problem because he has his university’s high-performance computers at his disposal. He returns from the forest with terabytes of data, which he brings to acoustic engineer Paul Roe. In Papua New Guinea, the team recorded nearly 1,300 hours of soundscapes. It takes Roe less than a day to translate Game’s files into usable numerical data using a script he created to process very large datasets.
But is it too late to start bioacoustic mapping since so much biodiversity has already been lost? Where others find despair, Game sees hope: “Soundscapes aren’t just a way to diligently record the defaunation of our natural world; they are a tool in our quest to do something about the loss of species.”
“You always have to have hope, to feel you have some kind of agency, or you will get depressed,” Game admits. “The forest faces enormous challenges, but I’m hopeful.”