In much of the world, nature sounds like birds or maybe insects — the cheerful, rhythmic cheeps of a robin, say, or the buzz of a cicada.
In Puerto Rico, it sounds like frogs. Lots and lots of frogs.
If you walk into pretty much any forest on the island you’ll hear a loud, almost deafening chorus of amphibians. The lead singer is the common coqui, a tannish treefrog with big, sticky toes. They are Puerto Rican icons no larger than a ping-pong ball that call — CO-KEE, CO-KEE — to attract mates and mark their territory.
Puerto Rico, a Caribbean island and US territory, is home to more than 15 native species of frogs. And these animals have a lot to say. The number and diversity of frogs, and their collective voice, can indicate the health of an ecosystem. Frogs, like other amphibians, tend to be sensitive to subtle changes in climate and water quality, making them reliable environmental sensors, according to Marconi Campos-Cerqueira, an ecologist.
When lots of frogs are singing, it’s a sign that the forest is healthy. A quiet forest is cause for concern.
Scientists are increasingly listening to their calls. In the last two decades, a large number of researchers have turned to bioacoustics, the study of all kinds of animal sounds, to monitor the health of the environment. It’s pretty simple: They place a few mics in the forest to gather sound, and then run the data they retrieve through computer software. That software reveals what animal species are there, and when.
For years now, scientists in Puerto Rico have been using bioacoustics to monitor environmental changes. By listening to the sounds of frogs, they’re figuring out how climate change — and the storms and heat waves that come with it — is altering life on the island.
On a cloudy morning earlier this year, I drove into the mountains of central Puerto Rico with Kris Harmon, a scientist at WildMon, a nonprofit that monitors ecosystem health in Puerto Rico and elsewhere. Parts of the road were crumbling, likely eroded from excessive rain. Climate change has been making hurricanes more destructive — including Hurricane Maria, which struck Puerto Rico in 2017 — in part by loading them up with more moisture.
Harmon parked above the clouds at the edge of a lush forest. Before I even opened the door I could hear frogs singing. CO-KEE! CO-KEE! There was also the more occasional KI-KI-KI-KI-KI from a red-eyed coqui, a small tree frog with rust-colored eyes.
“The frogs are omnipresent,” Harmon said.
We walked into the woods down a steep, overgrown path flanked by ferns, palms, and bromeliads (plants that look like the tops of pineapples). Harmon soon spotted what he was looking for: a microphone strapped to the trunk of a large tree.
This mic, housed in a simple black box, records forest sounds. It’s part of a long-term project, initiated a few years ago by Rainforest Connection, another environmental group, to monitor wildlife in Puerto Rico. There are some 29 other mics just like it in different habitats around the island that have been running for years.
The hardware is simple. The mic is attached to a circuit board and a battery that records one minute of sound every five minutes to a built-in memory card. Every few months, those cards are collected and the recordings are uploaded online and converted into spectrograms, which show sound data in visual form. Then, the software essentially scans the spectrograms for matches within an existing library of animal noises. The data can tell researchers which species are present at each recording site and, more broadly, how the range of sounds in one habitat differs from another.
The mic Harmon showed me, in a site called Toro Negro, has picked up the mutterings of a number of different animals in the last few years. These include those two noisy frogs, of course — the common and red-eyed coquis — as well as several other frogs, including melodius and grass coquis, that have their own distinct song. It has also detected a variety of birds, such the Puerto Rican tody, an emerald green species with bright red throats. (The software that analyzes sounds for this project only detects about 40 species of interest, Harmon said, including those that are threatened with extinction.)
The data isn’t perfect. Some animals, such as snakes, don’t call at all. Plus, sound alone can’t tell you much about abundance — i.e., how many of a certain species are found in a particular place.
“Discerning whether 100 detected calls are from one bird in mating season or dozens of less active individuals remains challenging,” said Ben Gottesman, a bioacoustics expert at the K. Lisa Yang Center for Conservation Bioacoustics at Cornell University. “Estimating population density through bioacoustics is still quite complex.”
But if your goal is to figure out what species live or travel through a particular area, Harmon says, bioacoustics technology is “pretty hard to beat.”
On a basic level, microphones reveal what does and does not live in a place, without having to spend a lot of time doing field work. This helps scientists easily create maps of biodiversity. Local authorities and nonprofits use these sorts of charts to establish new wildlife reserves.
But these sounds can also help researchers understand how ecosystems are changing. In a pair of studies based on bioacoustics, Campos-Cerqueira, the chief scientific officer at WildMon, who works with Harmon, found that a number of frog and bird species appear to be moving to higher elevations, where it’s cooler, likely due to rising temperatures. In lowland areas of the Luquillo Mountains, a rainforest in eastern Puerto Rico, for example, recording devices no longer detect a handful of frog species that once lived there.
“Many bird and frog species are changing their distribution on the island,” Campos-Cerqueira told me. “A lot of species are going up [higher] into the mountains to track more favorable temperature and humidity conditions.”
This movement of frogs and birds makes existing protected areas less useful — because unlike the animals they safeguard, they’re stuck in place. By mid-century, much of the habitat that’s cool and wet enough for frogs will likely fall outside of wildlife reserves, according to a 2021 study led by Campos-Cerqueira.
Other research has used mics to understand how superstorms affect wildlife on the island. In another 2021 study, led by Gottesman, scientists compared communities of birds and insects before and after Hurricane Maria using bioacoustics. Maria was the second most powerful storm to hit Puerto Rico in recorded history, killing nearly 3,000 people and damaging or flattening millions of trees.
Though not all that surprising, the authors found that sounds made by birds and insects (at least those that produce low and medium frequency noises) dropped by about 50 percent following the hurricane. It took one to two months for the sounds of insects to return to pre-storm levels, the authors found. The variety of bird calls took even longer to bounce back.
According to the analysis, audio recorders detected a drop in bird calls two days before the storm hit, adding evidence that some animals — from warblers to sharks — may be able to detect incoming storms days in advance and flee the area. “In this time of rapid environmental change,” Gottesman and his coauthors wrote, “soundscapes can signal the ways in which altered disturbance regimes are transforming ecosystems.”
Tropical ecosystems are complicated. There are literally thousands of species of plants, animals, and fungi that all influence each other, like a complex machine that no one fully understands.
In just a few days in Puerto Rico this year, I saw boa constrictors and lizards and waterbirds, lots of weird bugs, and so many frogs. They all play a role, including species that we’re only just discovering.
This complexity makes figuring out how ecosystems are changing — really, how we are changing them — extremely difficult. And that in turn makes it hard to hold those who are responsible for that change accountable.
More and more, major financial firms, including Fidelity International, increasingly want to know the impact of their investments on the environment, and how declining ecosystems put those investments at risk. How, for example, are insect pollinators faring and what impact might that have on agricultural commodities and producers?
Advances in bioacoustics are making it much easier to answer these sorts of questions. When paired with other tools that are now similarly mainstream — including motion-sensing cameras, satellite imagery, and eDNA analysis, where scientists analyze the bits of DNA in a habitat — audio recordings and the software to analyze them are ushering in a new era of wildlife monitoring.
This isn’t just about pointing out problems, such as the destruction of a food web. It’s also about helping monitor progress in restoring nature. A study published last year in Nature Communications, for example, showed that bioacoustics can be used to measure the recovery of wildlife in tropical forests that were previously cleared for farming and ranching.
The researchers analyzed audio from forests in northwestern Ecuador, finding significant changes in the community of noise-making animals in forests at different stages of recovery. “Clearly the regenerating forests are recovering species that are only really found in old growth forests, which is a really good sign,” Zuzana Buřivalová, a coauthor on the study, told my colleague Byrd Pinkerton earlier this year for an episode of the podcast Unexplainable.
As I spent more time in Puerto Rico, my ears became more discerning. Instead of just recognizing “a frog,” I was able to identify different species of coqui, even when I couldn’t see them. (They were nearly impossible to find, even when they seemed to be calling from just inches away.) In those subtle musical differences I was starting to detect — in that variety of frog croaks — are secrets to the health of these forests. To unlock them we just have to listen.
Clarification, June 25, 11:45 am: Ben Gottesman’s professional affiliation has been detailed further; he works at the K. Lisa Yang Center for Conservation Bioacoustics at Cornell University.
