“I officially started June 2019. I coded everything I needed to, but the criteria turned out to be wrong. Without the proper criteria for this data, the numbers mean nothing. So I had to go back and redo it. Then, it was July, and I broke my leg. I went home to recover while still coding with the new criteria. In October, we realized that the sampling efforts were wrong, and I had to code it all again. This is the third iteration.”
Simone Gatson ’20, an undergraduate researcher involved in the Elephant Listening Project (ELP) at Cornell, shows me her laptop screen. On it is a spectrogram with a series of black lines across the bottom. She tells me that these black lines are visual representations of elephant rumbles. At regular intervals above these lines are other lines. The bottom-most line is called the fundamental, and the lines above each fundamental are its harmonics. Sometimes multiple fundamentals overlap, and other fundamentals have greater or fewer harmonics above them.
To the untrained eye, however, it simply looks like a lot of lines. “This is the type of data that I’m working to categorize,” Gatson smiles. “You might be able to see why I’m on my third try.”
Coding and Classifying Elephant Sounds
The ELP, currently under the direction of Peter Wrege, is based at the Cornell Lab of Ornithology. Its team of researchers and data analysts is coding and classifying thousands of sound files collected from African forest elephants in Nouabalé-Ndoki National Park in the Republic of Congo. Sound detectors that are placed throughout the forest detect low-frequency sounds the African forest elephants produce, called rumbles.
“There are three species of elephants,” Gatson explains. “Savanna elephants, Asian elephants, and forest elephants.” Unlike the African savannah elephants that live in open areas and are easy to spot, the African forest elephants are a bit more elusive. “Forest elephants live under a dense canopy. They’re more difficult to observe. You really only see them when they come to meet in forest clearings, known as bais,” says Gatson.
Because the forest is so dense, researchers believe that the forest elephants produce low-frequency rumbles—sometimes at frequencies below what the human ear can detect—in order to communicate across long distances with each other. Through their rumbles, the elephants can send information to one another and coordinate meeting places at various bais in the forest.
These calls also help researchers estimate population size and density, habitat range, and location. Once they have ideas of the elephants’ locations, they can send that information to rangers in the park to inform anti-poaching efforts. “If we know where the elephants are,” Gatson says, “then we know where we should be focusing our efforts to safeguard them.”
Because forest elephants are harder to observe, they haven’t been studied as extensively as savanna elephants. As such, a lot of the ELP’s work is focused on setting up protocols and procedures to analyze, code, and categorize the sound data they’ve collected in a meaningful way. Having had to re-evaluate her data several times with refined criteria, Gatson says, “I’ve learned that it’s always three steps forward, two steps back. I’ve had to go back a lot, but it’s important to have a dataset that you feel good about.”
Though sometimes time-consuming, working in a new area of research has been exciting for Gatson. “I found this pattern in many of my calls where a lot of rumbles are overlapping with each other. I’ve taken to calling them rumble bouts.” And so far, that name has stuck. Gatson has had the opportunity not only to work with data, but she is developing protocols—and even terminology—for future researchers with the ELP.
“I found this pattern in many of my calls where a lot of rumbles are overlapping with each other. I’ve taken to calling them rumble bouts.”
For instance, the ELP hypothesizes that these calls may also be hiding a lot of additional information that’s waiting to be discovered. “So far, I’ve classified these rumble bouts into at least five different groups, based on a variety of characteristics. I look at things like how long each rumble lasts, how many elephants are calling at once, and how long the rumble bout goes on for,” Gatson says.
Others working on the ELP are trying to determine whether there’s hidden information about age or sex in the harmonics of the rumbles. “For example, it’s pretty easy to tell if the elephant is a baby. It’s quieter, higher frequency, overall just a little more pathetic-sounding,” Gatson laughs.
Women of Color Working with Animals
Although Gatson has loved working with animals since she was four years old, she came to Cornell with a certain perception. “I was thinking that I couldn’t do that, because I hadn’t seen a lot of people who looked like me actually doing it,” she says. “But I ended up taking the basic biology and chemistry courses anyway. I figured I might as well apply myself to doing the things I wanted to do. And I wanted to try and increase the visibility of women of color in science.”
In many ways, the ELP has enabled Gatson to work on a project that’s bigger than Cornell. “The idea of research can be scary because you might think you’ll be doing everything alone. But for me, the collaboration has been the best part,” she says. “I’ve met a group of kind and intelligent and really great people, and I’ve gotten the chance to work with them on an amazing project. I’m doing the things that I was scared of doing when I first came to Cornell.”
Gatson laughs when I ask her if she’s gone on an expedition to the Congo to see the elephants for herself; she’s still working on her insect phobia. For now, she’s happy to listen to the sounds of mysterious forest creatures more than half the world away—from the comfort of her desk.