On the island of Java, every gamelan orchestra tunes its bronze keys differently, so two neighboring ensembles can sound slightly distinct. Researchers used this quirk to turn climate data into music and test whether the chaotic nature of El Niño survives in sound.
El Niño is a fluctuation in Pacific sea surface temperatures that recurs irregularly, roughly every two to seven years. Such fluctuations are usually shown as graphs, but graphs are only seen with the eyes. Hearing is better at catching repetitions and rhythm, so the data were converted into notes. They took two Javanese scales: pelog, with uneven intervals between notes, and slendro, with more even spacing. Each note was assigned a loudness depending on how far the temperature deviated from the norm and how quickly it changed.
After that, the sound was split into frames and two features were measured: brightness (how high or low the pitch is) and energy (how loud). Then trajectories were built in the space of these two features, and they looked at how often a trajectory returns to places it has already visited. It turned out that when layers alternate, the trajectory often follows familiar paths, while when all layers sound simultaneously, it wanders over a large area. The scales also matter: in pelog, high notes are quieter, while in slendro, such a relationship is almost absent.
Previously, it was unclear whether a sonification preserves the dynamics of the data. Now there is a simple way to measure it: compute recurrence, area, and path length. This method could be useful, for instance, in climate communication, when you need to convey alarming changes to people who are not used to reading graphs.
Gamelan ensembles are not tuned to a single standard: each orchestra has its own microtonal tuning, called 'embet', making its sound unique. This study used an approximation of equal temperament, so a real Javanese gamelan would sound different.