Acoustic Arrays, or 'Who Said That?'

At right is an aerial image of the Dzanga Bai in Central African Republic, with the location of seven ARUs (Autonomous Recording Units) around the perimeter of the bai (yellow spots) and the location of the observation platform (brown rectangle). This was the setup in 2002 for Dr. Mya Thompson's Ph.D. research into forest elephant communication behavior.

Each of the seven ARUs was equipped with a GPS unit, which provided time information with an accuracy of milliseconds. This allows one to 'stitch' together the sound file from each of the units such that the timing is precisely aligned. It is then possible to analyze the difference in arrival time of the same elephant rumble on different ARUs, and triangulate to the location of the rumble source.

Below is a short animation based on the array used by Peter Wrege in 2007 at Langoué Bai, Ivindo National Park, Gabon. Although this sort of 'detective work', to find out which specific animal made which specific call, is very time consuming, the rewards in terms of our understanding of elephant communication are great.

Content on this page requires a newer version of Adobe Flash Player.

Get Adobe Flash player

Here is a real example of how this process lets us discover hidden interactions.
The figure below left shows six ARU recordings 'stitched' together. Six different elephant calls are indicated with colored boxes on one 'channel'. If you look very closely, you can see that the calls arrive on the different channels at very slightly different times (you might see this most clearly with the sixth call - it clearly arrived later on the 1st and 4th channels). This difference in arrival is what allows us to figure out who gave the call.

Of course, this 'stitching' and 'locating' happens long after the behavioral interactions occur. So while the recordings were being made, Mya and her team were simultaneously using video to film activities on the bai, and had a set of automatic still cameras taking a mosaic of pictures covering the entire clearing.

The figure above at right shows the detective work completed for the calls marked on the spectrogram. Three different animals were contributing to this segment of recording. We've colored the calls in the spectrogram according to which animal made them.

One of the most exciting things about this methodology is that it allows us to link the behavioral context and demography of the calling elephant to each recorded call. It is this linkage that will allow us to gradually decode the language of the forest elephant.

Related Topics

© Katy Payne

Koto and Melebu, Bayaka field assistants, operating the video camera on the observation platform at Dzanga Bai