Synchronized swimming for seal migrations

The research team analyzed and extracted features from raw GPS tracking data derived from a tagging program involving a total of 321 tagged animals tracked over a decade and mapped the data parameters to the sonic parameters of frequency (pitch) and amplitude (volume). The concept was that each animal traveling within a group would be assigned an instrument, a cello, with a different note, and that the distance from the colony would be coded as the pitch of the note. The spread of the group across the ocean will be coded as volume, with the group sounding loud when they were traveling closer together.

Combining music and movement is not unusual—but translating the movements of migrating marine animals into musical notes certainly is. An international research team including KAUST scientists have created a sound symphony using data charting the movements of northern elephant seals in the Pacific Ocean. This sonification technique provides surprising insights into group dynamics and synchronicity.  

“Many studies have analyzed single-animal tracks, but collective movement is rarely addressed,” says Carlos Duarte from KAUST, who led the project in collaboration with colleagues including Madhu Srinivasan from KAUST’s Visualization Core Lab, and scientists in the United States.

“When faced with more than a million data points capturing the underwater movements of 321 individual seals over 10 years, we needed a novel way of getting the data to 'talk’ to us,” Duarte explains.

A key challenge when faced with big data is finding meaningful ways of representing the data to extract valuable information. Sonification—adding sound to data to trace individual movements—has a long history within scientific disciplines, but has never been used on animal migration data before. The team spent a lot of time deciding which elements of the data could be represented using sound.

“We only had raw, time-stamped GPS traces for the seals’ migration to begin with,” says Srinivasan. “We explored many parameters derived from these traces, such as speed, straightness-index, and group dispersion. In the end, we focused on the position and spread of individual groups (comprising between four and 30 animals) relative to the location of the main colony.”

Read the full article