The Acoustic Search

In the field

ARU mounted on a tree (left) with its battery (right). Photo by Chris Tessaglia-Hymes

To search for acoustic evidence for Ivory-billed Woodpeckers in Arkansas and other states within the historical range, we record ambient sounds using autonomous recording units (ARUs). ARUs are programmable, battery-operated digital audio recorders developed by the Cornell Lab of Ornithology’s Bioacoustics Research Program. Each ARU contains a microprocessor, 12-bit analog-to-digital converter, an omnidirectional microphone, preamplifier and signal conditioning circuitry, and a hard disk for storing audio data. These components are packaged in a cylindrical PVC housing, and attached to tree trunks two to three meters above the ground or water surface. ARUs are typically deployed for periods of two to four weeks.


ARU in Arkansas. Photo by Chris Tessaglia-Hymes

ARUs are programmed to record for two four-hour periods each day, the first beginning 30 to 45 minutes before sunrise, the second ending 30 to 45 minutes after sunset. The range at which an ARU could detect sounds of an Ivory-billed Woodpecker is unknown, because there are no data available on the volume of kent calls or double knocks. We estimate, however, that these signals would be detectable by ARUs up to distances of approximately 200 meters.

We select recording sites based on habitat quality, locations of previous Ivory-billed Woodpecker sighting reports, and presence of possible ivory-bill roost/nest cavities and feeding signs.

 

 

 

Reviewing and analyzing the sounds

Since the start of large-scale acoustic search efforts in 2004, our protocols for reviewing and evaluating ARU recordings have evolved in order to provide more consistent and informative evaluations of ivory-bill-like sounds. Our current protocol is summarized here.

To find sounds similar to those of Ivory-billed Woodpeckers in the ARU recordings, we use a multi-step process:




1. Automated screening by computer:  The digital recordings are scanned by software that detects sounds similar to known vocalizations of Ivory-billed Woodpeckers (from the 1935 Allen-Kellogg recording), and to double-knocks from other Campephilus woodpeckers.



2. Initial human screening:  An acoustic analyst reviews all of the computer’s detections. Most of the sounds flagged by the computer are easily discarded at this stage as not being similar enough to ivory-bill sounds to warrant further attention. The computer flags many “false alarm” events because we adjust the software to be very sensitive, reducing the chance that a real ivory-bill call might be missed. Sounds that pass this stage are forwarded to the next stage of review.



3. Expert panel review:  A panel of three or more experts (outside of the acoustic analysis team) reviews all of the sounds that pass stage two. The expert panel categorizes each sound as “implausible” or “plausible.” “Plausible” events are further categorized depending on whether a potential alternate source is identified, and if that alternate source is positively identified elsewhere on the deployment. Sounds categorized as “implausible” are either positively identified as an alternate source, or are deemed to be too different than an ivory-bill.


Plausible categories are:

  • P1: Plausible Ivory-billed Woodpecker, no likely alternative known
  • P2: Plausible Ivory-billed Woodpecker, alternate possibility identified but not present in recording
  • P3: Plausible Ivory-billed Woodpecker, alternate possibility identified and present
  • P4: Insufficient signal for full analysis

“Plausible” sounds are scored on various criteria, receiving a point for each positive response to one of several questions. A higher score indicates a greater likelihood that the sound originated from an Ivory-billed Woodpecker.

Scoring criteria for vocalizations:

1.  Is the harmonic interval between 580 and 780 Hz?

2.  Is harmonic emphasis appropriate?

3.  Is the event part of a biologically appropriate series?

4.  Is there a temporal context or co-occurrence with other events of interest on the same day?

5.  Is there a clear temporal context or co-occurrence with other events of interest across days?

Scoring criteria for double-knocks:

1.  Is the inter-knock interval between 60 and 120 milliseconds?   

2.  Is sound resonant and woody?

3.  Is there an absence of confounding woodpeckers?

4.  Is the event part of a biologically appropriate series?

5.  Is there a clear temporal context or co-occurrence with other events of interest on the same day?

6.  Is there a clear temporal context or co-occurrence with other events of interest across days?

At every stage of the review process, researchers compared suspect sounds not only with those of ivory-bills and other Campephilus woodpeckers, but also with a variety of similar sounds from other species, and carefully considered the surrounding context.

What have we discovered so far?

Here we present some examples of “plausible” sounds collected in the Big Woods of Arkansas. Note: this website is not intended to be a complete and final analysis of our acoustic monitoring and research. Rather, we aim to provide a sampling of sounds that we believe are suggestive of ivory-bill and a number of “sound-alikes” that we hope will help inform other searchers about what to listen for. We are presently working on peer reviewed publication that will explain our findings in detail.

Next page

Previous page