Spatiotemporal and probability variations of surface PM_2.5 over China between 2013 and 2019 and the associated changes in health risks: An integrative observation and model analysis

We used statistical methods and the GEOS-Chem model to interpret the observed spatiotemporal and probability variations of surface PM_2.5 concentrations in China from December 2013 to November 2019, as well as to assess the drivers for the variations and the implications for health risks associated with long-term and short-term exposure to PM_2.5. Annual and seasonal PM_2.5 concentrations have decreased over most areas in China during the 6-year period. We decomposed the observed day-to-day variation of PM_2.5 concentrations in eastern Chinese cities and found that it showed two distinct major spatial modes, which fluctuated in strength seasonally. The first mode, characterized by most of Eastern China being in the same phase, was mainly associated with the regional ventilation of pollutants. The second mode showed a dipolar pattern between the Beijing-Tianjin-Hebei area and the Yangtze River Delta area and was more prominent in summer. Using model simulations, we showed that this dipole mode was chemically driven by the secondary formation of sulfate in summer. We further used a gamma distribution to succinctly interpret the changes in the probability distributions of PM_2.5. We found that the nationwide decline in seasonal mean PM_2.5 concentrations mainly reflected decreased occurrences of extremely high PM_2.5 concentrations, which was strongly driven by the interannual variation of meteorology. These changes in the annual means and probability distributions of PM_2.5 since December 2013 has led to significant decline of the estimated mortality risks associated with long-term and short-term PM_2.5-exposures. Regions that were less polluted saw the largest relative benefit per unit decrease in PM_2.5 concentration, due to the steepness of the exposure-response curve at the low-concentration end. Our integrated methodology effectively diagnosed the drivers of PM_2.5 variability and the associated health risks and can be used as part of the decision tool for PM_2.5 pollution management over China.