On October 10, the day after massive wildfires began in Northern California and a thick haze of smoke spread around the region, two Google Street View cars began driving on local roads. On top of each car, a mobile air-quality system sucked in outdoor air and analyzed the level of pollution–at times, the worst air quality on record for the Bay Area.
If you live in San Francisco or Oakland or Santa Rosa, you might have checked a government website that posts estimates of the Air Quality Index for a handful of locations that have sensors. (I refreshed it obsessively to try to figure out when to go outside.) But the system used on the cars–which could eventually be deployed at a large scale after future fires, or for everyday pollution–makes it possible to get much more detailed data.
When the cars drove in the Berkeley Hills, for example, the data showed that the pollution was much worse than in the flat part of the city. “This is something that would not be reflected by a regulatory monitoring network, but provides important information for those living in those different regions and elevations as to the immediate impact of the fire emissions on their health,” says Melissa Lunden, chief scientist at Aclima, the company that designed the air quality platform and partnered with Google to test it on the cars.
The cars can also respond to changing conditions. For four days, the team designed a specific route to drive based on the direction of the wind, and to capture specific impacts of the smoke. On two days, the cars drove near the fires–in areas where it was safe to drive, and where they wouldn’t interfere with fire crews or other emergency responders–noting how rapidly pollution levels changed even over distances as short as one mile.
“You can see in our animated maps that we were able to capture more granularity,” says Lunden. “These initial takeaways help us better understand the value of hyper-local air quality and serves as an opportunity to understand how this data could inform better decision-making and action.”
On a drive up Highway 101 to the town of Sebastopol, for example, on October 10, the pollution varied between moderate and unhealthy over short distances. When the cars passed near government pollution monitoring stations, the data was similar. But the cars were able to also gather data in all of the areas that traditional stations don’t cover.
“The advantage of our sensing platform on the Google Street View cars is that it’s mobile,” she says. “The flexibility of the mobile platform allows us to respond to the shifting conditions of the day. We can get close to a source of air pollution, but we can also explore the effect of the plume in downwind regions, such as tracking regions of concern and areas that are less affected at that time. This real-time data could allow residents to plan (or postpone) outdoor activities.”
For now, the system is still being piloted with only two cars. But Aclima envisions that it will eventually have a network installed on large fleets of cars and city buses, along with building interiors and outdoor stations, streaming billions of data points to help people understand, in detail, how polluted the air they’re breathing is.