Listening for Ivory-bills, the High-Tech Way

Lab researchers seek audible clues from the Big Woods

By Russ Charif

In James Tanner's landmark 1942 book, The Ivory-billed Woodpecker, he observed that the best way to find these birds in their dense forest haunts is to listen. Indeed, in nearly three years of field work, Tanner always found an Ivory-billed Woodpecker first by hearing it, then by going toward the sound.

So in 2004, when the Cornell Laboratory of Ornithology began to search for ivory-bills in the Big Woods of Arkansas, listening had to be part of the plan. But all the sharp-eared birders that could possibly be recruited for this secret search could barely scratch the surface of the vast area to be covered. And if a searcher in the field heard a faint call, like the far-off toot of a toy trumpet, the sound would be gone in a heartbeat, leaving the searcher to wonder--was that the nasal kent call of an ivory-bill or another bird in the distance with a similar-sounding call?

An autonomous
recording unit
(ARU)

Photo by Susan Spear/CLO

The solution was to deploy electronic listening devices called autonomous recording units (ARUs) at dozens of sites throughout the search area. ARUs could listen for weeks at a time, recording every sound for careful analysis later. During the search for the ivory-bill in 2004-05, researchers from the Lab of Ornithology used ARUs to gather more than 17,000 hours of recordings from 153 sites across half a million acres of potential ivory-bill habitat. The goal of this large-scale acoustic survey was to give us a clearer idea of where ivory-bills were spending time, and thus to help focus current and future research into the numbers and status of the species.

We have spent thousands of hours of computer and human time searching through these recordings for the distinctive nasal kent calls and double-knock display drums of Ivory-billed Woodpeckers, and there's still more to be done. The effort has involved field biologists, engineers, computer programmers, sound analysts, and archivists in a collaboration that has stretched the limits of our ability to search large areas for rare species. Along the way, we have had to invent new technologies and refine existing methods.

A portion of the study site, from above. Autonomous recording units helped monitor vast areas.

Photo courtesy of USGS and TerraServer

From December 2004 through May 2005, a two-person field team worked full-time deploying and retrieving 24 ARUs at sites throughout the Big Woods search area, sometimes paddling or hiking through chest-deep water for hours to reach remote sites (see page 13). The searchers chose sites that seemed promising for ivory-bill encounters, based on habitat quality, locations of previous sightings, and possible signs of ivory-bill foraging activity (large areas of bark stripped from dead or dying trees).

Each ARU, packaged in an 18-inch-long plastic pipe, was tied to a tree, along with a battery pack to power the unit. A microphone protruded from the bottom end of the pipe, covered with a furry windscreen. The ARUs were programmed to record for four hours in the early morning and four hours in the late afternoon and evening, times when ivory-bills were most active, according to Tanner.

Two to four weeks after the field team deployed each ARU, they retrieved the unit, downloaded the data, and sent an external hard disk back to the Lab of Ornithology for analysis. From January to May, around 900 hours of recordings were collected every week!

Back at the Lab, the challenge was how to dig through this avalanche of data for what would undoubtedly be very rare occurrences of either of the two distinctive sounds that ivory-bills were known to make. The ivory-bill's nasal tooting call is reminiscent of a White-breasted Nuthatch's call, but more powerful. The only known recording of this sound, about five minutes long, was made in the Singer Tract in Louisiana during the historic 1935 expedition by the Lab's founder, Arthur "Doc" Allen, sound recording pioneer Peter Paul Kellogg, and then-graduate student James Tanner.

The ivory-bill?s other distinctive sound was never definitively recorded, and is known only from written descriptions: a loud rapid double knock, BAM-bam!, used as a long-distance contact signal.

This double-knock signal is shared by most members of the genus Campephilus, large woodpeckers distributed throughout the tropics in Central and South America. Since the double knocks of all of these species are nearly identical to each other, we assume that the ivory-bill?s sound was very similar.

Listening to all of the ARU recordings was not practical--doing so would take years. But somewhere buried in those thousands of hours could be a few seconds containing crucial information on the presence of ivory-bills. To speed up the search process, researchers rely on advanced sound-processing software developed in the Lab's Bioacoustics Research Program.

A program called XBAT (eXtensible BioAcoustic Tool) scans the recordings for sounds that match template calls extracted from the 1935 recordings and from recordings of double knocks made by other Campephilus woodpeckers. XBAT can scan 50 to 100 hours of recorded sound in a single hour, keeping a log of every sound that bears any resemblance to the templates.

A computer-generated sonogram shows six kent calls from the only definitive recording of an Ivory-billed Woodpecker in 1935. These calls are the standard with which similar sounds from the Big Woods are compared.

To minimize the chance of overlooking a faint or aberrant sound, the program is configured to be very sensitive, even though this results in numerous false alarms. Analysts then use XBAT to review the logs, examining visual images of the sound, called spectrograms. In many cases, mistaken detections can be quickly and easily recognized this way. When analysts find a logged sound that they can't visually rule out as an ivory-bill, they listen to it, comparing it to known recordings of ivory-bills and other species that make similar sounds.

Our bioacoustics team has reviewed tens of thousands of candidate sounds, and several thousand hours of recordings still remain to be processed. Among the sounds we've reviewed, we've found several isolated double knocks and one short series of double knocks that are extremely similar to those of other Campephilus woodpeckers (see page 14). So far, we have not found long series of recorded double knocks like those heard by observers on a couple of occasions during the 2004-05 search.

We have also found a handful of sounds that bear a striking resemblance to distant ivory-bill kent calls. A few expert birders who have listened to these sounds have suggested that they might have been made by Blue Jays, which are abundant at the recording site, and are notorious for the variety of their vocalizations and for their ability to mimic other species (usually predators such as Red-shouldered Hawks). Other experts disagree that they could have been made by Blue Jays. We are currently working on more detailed acoustic measurements and statistical comparisons between these kent-like sounds and sounds of Blue Jays and other possible sources to try to resolve the identity of these sounds.

Over the next several months, a small team of people wearing headphones will continue to sit in a windowless room, sifting through the sounds of the Arkansas swamp--the songbirds, the woodpeckers, the owls, the ducks, the coyotes, the rain--listening for the sounds of North America's rarest bird. The results of their efforts will help steer the direction of future search and conservation efforts in the Big Woods region. As the quest for ivory-bills in Arkansas and elsewhere continues, we envision a day when our children and grandchildren may hear the sounds of living Ivory-billed Woodpeckers echoing through the great southeastern swamp forests.

Russ Charif is a research biologist in the Bioacoustics Reserch Program. He is the acoustics leader in the search for the Ivory-billed Woodpecker.

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