91av

Air pollution in China and Taiwan is estimated to have resulted in the premature deaths of 30.8 million adults since 2000.

Yang Liu at Emory University in Atlanta, Georgia, and his colleagues used satellite imagery to quantify the amount of air pollution over mainland China, Hong Kong, Macao and Taiwan between 2000 and 2016. The team used imagery taken by NASA satellites to estimate the concentrations of PM2.5 – particulate matter less than 2.5 micrometres in diameter.

One measurement the satellite takes is the amount of sunlight that has been scattered or absorbed by particles in the air. Combining these readings with PM2.5 measurements from ground monitoring stations, as well as information including meteorological conditions and road networks, the researchers trained a machine-learning algorithm that was able to predict cumulative PM2.5 exposure over a 17-year period.

To estimate the total mortality linked to air pollution, they then used historical data from a study of 116,821 adults in 15 Chinese provinces, which quantified the link between long-term PM2.5 exposure and non-accidental death. The team found a roughly linear relationship between PM2.5 exposure and mortality, up to a certain point.

“The people who live in the most polluted regions get disproportionally harmed,” says Liu.

The highest per-capita deaths due to air pollution were in the north-eastern provinces of Hebei, Henan, Shandong and Tianjin.

To date, most air pollution monitoring has been done from stations on the ground. In China, these are concentrated in urban areas, which doesn’t account for some 600 million people who live in rural areas. In addition, measurements before 2013 are scarce.

The advantage of using satellite imagery to determine PM2.5 levels is that it is more comprehensive and also provides an estimate of historical air pollution, says Liu.