Deep Listening

New tools open up the world of animal sounds

By Molly Brewton and Miyoko Chu

Chris Clark listens to sounds of fin whales from a microphone array towed by a boat in Baja California. The research showed that only male fin whales sing loud songs, possibly to communicate with females about the location of good patches of food.

Jane Moon Clark

The idea clicked when Chris Clark, an engineer and biologist, was standing on a cliff overlooking the ocean. Every now and then, he saw a whale surface then disappear beneath the water, a hidden realm filled with the voices of whales. At the time, in the 1970s, researchers were listening to whales with underwater microphones, called hydrophones, but they couldn't tell who was saying what. To solve this problem, Chris envisioned dropping an entire array of hydrophones into the water, connected by cables to a computer. Based on how long it took each sound to reach the hydrophones, the computer could calculate where the sounds had come from.

It was the first of numerous innovations that would enable Chris and other researchers around the world to explore the three-dimensional soundscape of wild animals. Chris, now director of the Cornell Lab of Ornithology's Bioacoustics Research Program, has also led a team of scientists, computer programmers, and engineers in adapting, inventing, and proliferating the tools to study what we could not otherwise hear or comprehend--the infrasonic vocalizations of forest elephants, the voices of thousands of individual birds migrating through the night sky, the intricate meaning of a dawn chorus, the presence of some of the rarest species on earth, including the Ivory-billed Woodpecker.

Sounds in the Sea

In 1987, Chris packed up his hydrophones and headed for the coast of Alaska to listen for endangered bowhead whales. Conservation policies relied on estimates of how many whales there were, but efforts had been hampered because the whales could only be counted where they surfaced in patches of open water between ice.

Chris's crew extended their senses into the sea by dangling hydrophones into the water through ice holes. As they sat in the "sled shed"--a hut atop a snowmobile--they could hear, through headphones, the moaning vocalizations of bowhead whales. For six weeks, the hydrophones recorded the voices of migrating whales, monitoring an area 40 times larger than would have been possible visually. To make sense of the data, programmers created analytical software to locate and track the whales.

"We found that as many as 10 whales were coordinating their movements across 10 square miles," Chris said, "much like a flock of geese in the spring." The acoustic census also revealed that there were more bowhead whales than the visual counts had estimated; the species was recovering, more than 80 years after a ban on commercial whaling.

The study of whale vocalizations took another technological leap in the 1990s, when engineers at the Lab of Ornithology devised autonomous recording units (ARUs) that could be dropped to the bottom of the ocean for months at a time and retrieved later, along with all of the recorded data. Housed in a 17-inch glass sphere covered in a yellow plastic shell, the unit included a hydrophone, microprocessor, hard disk, and batteries. The researchers called them "pop-ups" because a signal could be sent to apply voltage to the anchor wire, causing the wire to corrode, break, and bring the unit popping up to the surface.

The guts of a ?pop-up,? designed by the Lab?s engineers to gather sounds from the ocean autonomously.

Christopher T. Tessaglia-Hymes/CLO

Chris and collaborators have used ARUs in oceans around the world to monitor whale populations and learn about their behaviors, including the challenges they face while communicating in an underwater world polluted by noise from shipping, oil drilling, military sonar, and other activities. The researchers are also devising microphone arrays that can be coupled with cellular phone or satellite technology to track movements of endangered northern right whales and relay the information to ships in real time, helping prevent deadly collisions.

From Sea to Sky

The Lab's bioacoustics team has also repackaged ARUs to record sounds from land and air. In the Central African Republic, Katy Payne and collaborators strapped ARUs onto trees--truck batteries and all--to record forest elephant vocalizations, many of which are so low in frequency that humans cannot hear them. Using video cameras, the researchers also filmed the elephants as they gathered in a muddy clearing, attracted by the salty minerals in the water and dust.

Using the films and recordings from the microphone arrays, the team is piecing together an extraordinary record of how the elephants communicated as they greeted one another, fought, bathed, and mated. By quantifying the number of vocalizations and the number of elephants, they are also establishing protocols to monitor this endangered species using sound--an important breakthrough since it is difficult to census these elephants by sight in their densely forested habitat.

At Fort Hood, in Texas, Kurt Fristrup and collaborators are using airborne ARUs to monitor two more endangered species: Black-capped Vireos and Golden-cheeked Warblers. They invented a system using weather balloons, radio-controlled circuitry, and a GPS navigation module to loft and guide ARUs about the size of a box of cereal high above the earth, enabling researchers to listen for the birds above terrain that is inaccessible because of unexploded ordnance from military training.

ARUs are also ideal listening devices for animals that are rare and difficult to find. ARUs are unobtrusive and endlessly patient, and they are immune to the discomforts of foul weather, mosquitoes, poisonous snakes, hunger, and fatigue. This has proved to be an advantage in the swamps of Arkansas, where one or more Ivory-billed Woodpeckers were seen in 2004 and 2005. Using ARUs, the Lab's search team has gathered more than 17,000 hours of recordings across half-a- million acres. With analytical software designed and created at the Lab over the years, as well as new programs devised specifically for the ivory-bill search, technicians have pinpointed knocking sounds and calls that could have been made by ivory-bills. The information is being used to help focus on the most promising areas to search.

On an even larger scale, Lab researchers are also working on a project to predict the nocturnal migration routes of birds. They have deployed ARUs from the Canadian border to the southern end of the Chesapeake Bay--an unprecedented geographic span. The recordings will help show the numbers and kinds of birds stopping over during migration.

Microphone arrays are also helping researchers unravel complex vocal interactions within a neighborhood of territorial male birds. In Costa Rica, Sandra Vehrencamp and collaborators tracked the voices of Banded Wrens, leading to the clearest evidence yet that birds communicate with one another in a network during the dawn chorus, each bird listening to all its neighbors simultaneously and adjusting its own song from moment to moment depending on what it hears.

Making Sense of the Sounds

With the ability to record thousands of hours of sounds, one of the greatest challenges is how to analyze it all. In the 1980s, computer programmers at the Lab created Canary software so that researchers could load recordings into a computer and generate pictures of the sounds, called spectrograms, on the screen. This made it much quicker to identify and quantify sounds than using the traditional method of printing spectrograms on paper from a machine and making measurements with pens and rulers. Since then, Canary has become one of the word's most widely used software tools for sound analysis. With Canary's successor, Raven, researchers can take their laptops right into the field and view spectrograms in real time, as the computer records the sounds.

The Lab's programmers have also developed a software program, called XBAT, that can scan recordings for sounds that match a template of an animal's call. For example, a recording of an Ivory-billed Woodpecker by Lab founder Arthur Allen in 1935 was used as the template for XBAT to search the recordings from Arkansas. By flagging sounds with similar characteristics, the program made it possible for researchers to find hundreds of possible ivory-bill calls in a matter of months, rather than the years it would have taken if they had listened to every second of the tapes.

Three decades after Chris Clark first pondered the problem of how to listen to whales, he continues to lead the Bioacoustics Research Program in engineering the tools to perceive wildlife in new ways. "Just as binoculars are critical for visual observations, advanced acoustic tools provide vistas into the natural world that allow us to answer questions and gain understandings that were previously unobtainable," he said. "Acoustic arrays combined with automatic detection software are telescopes that allow us to look deep into the night sky, literally and metaphorically. When combined with the Lab's expertise in birds, this allows us to describe and understand bird populations over spatial and temporal scales that are relevant both ecologically and for conservation.

"I love listening to the world. It's a spiritual awakening to listen to the songs of whales, the dawn chorus of birds, the calling of spring peepers," Chris said. "Paying attention to the complexities helps our understanding of the living world as exciting, vibrant, and important."

Molly Brewton is program assistant for the Lab's Bird Population Studies. Miyoko Chu is editor of BirdScope.

For permission to reprint all or part of this article, please contact Laura Erickson, editor, Cornell Lab of Ornithology, 159 Sapsucker Woods Rd., Ithaca, NY, 14850. Phone: (607) 254-1114. email: lle24@cornell.edu