Chapter 6 of 6
The air gets worse when it gets hot
Four years of hourly air-quality data over Atlanta, 1437 days joined against the same period's temperature record: ground-level ozone tracks heat strongly (r = 0.75). Fine particulate matter (PM2.5) barely moves with temperature at all (r = 0.02, the same daily-max basis as the chart below). That split isn't noise. Ozone forms photochemically, and heat and sunlight accelerate the reaction. PM2.5 in Atlanta comes mostly from transported smoke and regional haze, which doesn't care what the thermometer says.
Same days, PM2.5 instead of ozone
- Mean daily max ozone, coolest third of days
- 80 µg/m³
- Mean daily max ozone, hottest third of days
- 139 µg/m³, up 74%
- Highest ozone day on record
- 209 µg/m³ on 2025-06-26, 96°F that day
This chapter runs on CAMS, a 45-kilometer atmospheric model, not ground monitoring stations, the same coarseness tradeoff as the ERA5 weather record, for the same reason: one grid cell covers roughly the whole metro, with no genuine neighborhood-level resolution. A real station-level upgrade (OpenAQ's ground network) is scoped but not yet built. It needs an API key decision this site hasn't made yet, not more data-hunting. See Methods for the honest version of that gap, and About for how this data was pulled.