Every day of measured air quality in 29 American cities, one tile per day. The Clean Air Act's effect is not a statistic here. It is the color draining out of the grid.
Each dot is a city, colored by its typical air last year. Select one to open its record.
Los Angeles County, CA
Same scale and years throughout. Select any chart to open it.
Days each year the air was unhealthy for at least sensitive groups. Years still in progress are left out. Dashed lines mark national milestones, listed below the grid.
The numbers behind the grid. Check days measured before comparing a year.
| Year | Days measured | Median AQI | Worst day | Days >100 | Days >150 | Usual culprit |
|---|
What 46 years of monitoring can and cannot settle
The air got cleaner. Proving why is harder, but the record carries one piece of evidence that is difficult to explain any other way: the pollutants that vanished are precisely the ones the law was written against.
Pooling all 29 cities and asking what was responsible for each unhealthy day. Grouped by source, because that is what a law can act on:
Fine particulate was not measured anywhere before 1999, so the first bar has no particulate segment. That is an absence of instruments, not of dust.
| Cause of an unhealthy day |
|---|
These are shares of a shrinking total, not counts.
Carbon monoxide is almost entirely vehicle exhaust, and it stopped setting the daily index anywhere. Nitrogen dioxide, also combustion, went the same way. Catalytic converters, tailpipe limits and reformulated fuel were aimed at exactly these two. It dominates what is left because it is the hardest to remove: it forms in sunlight from precursors, rather than coming out of a pipe.
Fine particulate has no figure before 1999 because it was not measured. Its later fall, from a quarter of bad days to a sixth, should be read cautiously, since wildfire smoke makes it the most volatile line in the table from one five-year window to the next.
Ranked by unhealthy days removed from a typical year, comparing the early 1980s with the last five years. Percentage alone would flatter cities that had little to fix, so the grouping uses where a city ended up as well as how far it came.
Daily index readings show the outcome, not the mechanism. Fleet turnover, deindustrialisation and economic change all ran alongside the legislation, and because vehicles are replaced gradually no law leaves a sharp step in the data to point at. Two further things make the improvement look smaller than it was: fine particulate is missing from every year before 1999, and the ozone standard was tightened in 2015, so identical air scores worse now than it would have then.
One fact from outside this dataset carries most of the weight. Los Angeles County held about 7.5 million people in 1980 and roughly 10 million by 2020, and driving grew faster still. Air that improves while the activity producing the pollution expands is a considerably stronger result than air that improves on its own.
Each tile is one day, colored by that day's peak Air Quality Index across every monitor in the city's county. The index reports whichever pollutant scored worst, so a red day in 1985 is almost certainly ozone while a red day in 2021 is more likely wildfire smoke.
1980 onward uses the AQI values EPA publishes directly. 1976 to 1979 is available for Los Angeles only, reconstructed here from raw monitor readings using the current breakpoints, because EPA's published series does not reach back that far. Other cities show those four years as gaps.
Treat 1977 and 1978 with suspicion. Los Angeles County had just one working ozone monitor in 1977 and only partial coverage in 1978. Ozone peaks inland around Azusa and Glendora, so a thin network misses the worst air and those two rows read cleaner than the city actually was. 1976 and 1979 had adequate coverage and line up with 1980.
Fine particulate was not monitored anywhere until 1999. It now drives a large share of unhealthy days, so earlier years are measuring a narrower thing and the improvement shown here is, if anything, understated.
The list runs the twenty largest cities by population, plus nine chosen for what their records show: Riverside, which now has worse air than Los Angeles because it sits downwind; Bakersfield and Fresno in the San Joaquin Valley; Salt Lake City, whose winter inversions make January worse than August; and the Northwest cities where wildfire smoke arrives as sudden spikes rather than more bad days. Counties stand in for cities, so New York is Manhattan only and the others use the county containing city hall. Source: EPA AirData and the EPA Air Quality System API.