The sound of pollution
Artists like British “grime writer” Moose, who scrubs designs into filthy, smog-charred city surfaces (including the Broadway tunnel in San Francisco), have found novel ways to visualize air pollution for passersby. But now it’s also possible to experience air pollution with a different sense: hearing.
Using mass spectrometry, which helps scientists pinpoint the exact compounds present in an air quality sample, researchers from the University of California Berkeley have created a set of sound clips -- some of which, with their otherworldly blip, glitch and drone tones could be at home on an ambient dub track -- describing the chemicals floating in the air at various locations in California.
Here’s what the air in the heavily trafficked Caldecott tunnel, which connects Oakland with Conta Costra County, sounds like. Hydrocarbons, spat out of tailpipes, are responsible for the tones that characterize this clip.
In contrast, this is the soundtrack from a pine forest in the Sierra Mountains. The sounds in the beginning, which the researchers describe as “bubbly,” indicate the presence of volatile organic compounds, which trees produce to attract helpful insects and ward off pests. The sounds of emissions, blown up from Sacramento, make an appearance toward the end of the clip.
And this is what the air around Bakersfield, in Kern County, which has the unfortunate honor of being the city with the most “particle pollution” in the country, sounds like.
To find out more about the project, I interviewed Gabriel Isaacman, a graduate student with the research group that created the air quality sound files. Our conversation took place over email since he is in China for a conference
High country News: Air quality data is not usually something one can hear. Why did you decide to turn this data into sound?
Isaacman: It’s very important to explore and present data in new ways. Air quality is a subject we talk about a lot, especially in California, but most people don’t typically think too much about the science behind it or how we know what we know. Furthermore, most people, even those interested in science, will never have the opportunity to work with data or maybe even step foot in a lab, so I wanted to bring this data to the public in a way that is interesting and accessible. Making science more hands-on, or in this case ears-on, helps engage people and hopefully inspires people to learn more or maybe pursue it a bit deeper.
HCN: The sounds of pollutants like hydrocarbons from tailpipe emissions seem darker than the sounds assigned to natural chemicals given off from pine trees in the Sierras. How did you choose which sounds to assign to different chemicals?
Isaacman: It’s interesting to me that that is the case, because the tones were actually not assigned subjectively at all. I directly assigned tones based on the data. The way these files were made is that samples of air were collected on filters, and then analyzed in our lab using some advanced tools to separate all the compounds. What comes out of our instruments is information about the compounds in the (air). We identify (the compounds using a mass spectrometor and they are assigned) a set of identifying numbers, more-or-less unique for each compound. Compounds that fall apart more in our instrument end up with lower numbers.
I calculated a tone based on the numbers in this mass spectrum, approximately spanning the range of a standard piano. (The compounds get assigned) numbers in the range of (approximately) 30-500, and I mathematically turned these into frequencies from the range of human hearing. Compounds that have lower numbers identifying them happen to be hydrocarbons, making low chords, while the oxygenated compounds we measure in more remote environments tend to be higher and more diverse, leading to an interesting combination of drones and ‘plinks.’
HCN: In a recent article for the Atlantic, you and the writer Aaron Reuben mention that there are inaccuracies in air pollution models, chiefly due to the interaction of human-caused and natural emissions. How do these interactions lead to inaccuracies in pollution models?
Isaacman: Particulate matter, in general, is one of the largest sources of uncertainty in climate models, and it turns out we have historically not been great at modeling it, usually drastically underestimating the amount of particulate matter out there. As our field has gotten more advanced, of course, our models have gotten better, so more and more we are able to predict the amount of particulate matter. But often our predictions of its chemical properties then suffer, which is an important aspect in understanding climate and health impacts.
There is some evidence that a majority of particulate matter actually comes from chemical interactions between human and natural emissions, but we still have limited knowledge about this chemistry. Some evidence suggests human emissions increase formation of particulate matter from forest emissions, but it is still an area of active research.
Every type of emission, (from) pine trees, diesel vehicles, meat cooking etc., releases a unique combination of compounds that give it its unique smell, so in a sense what we are trying to do is not to (simply) hear the differences, but to smell the differences between different particulate matter. A detailed understanding of the compounds in the particles gives us an idea of where they came from and how they were formed. Investigating this through sound gives us a new way to think about the data.
Brendon Bosworth is a High Country News intern.
Sounds of smog files courtesy Gabriel Isaacman/Aaron Reuben.
Image courtesy wikimedia user Andreas 06.