The Influence of Dynamics and Emissions Changes on China’s Wintertime Haze

Abstract Haze days induced by aerosol pollution in North and East China have posed a persistent and growing problem over the past few decades. These events are particularly threatening to densely populated cities such as Beijing. While the sources of this pollution are predominantly anthropogenic, natural climate variations may also play a role in allowing for atmospheric conditions conducive to formation of severe haze episodes over populated areas. Here, an investigation is conducted into the effects of changes in global dynamics and emissions on air quality in China’s polluted regions using 35 simulations developed from the Community Earth Systems Model Large Ensemble (CESM LENS) run over the period 1920–2100. It is shown that internal variability significantly modulates aerosol optical depth (AOD) over China; it takes roughly a decade for the forced response to balance the effects from internal variability even in China’s most polluted regions. Random forest regressions are used to accurately model (R2 > 0.9) wintertime AOD using just climate oscillations, the month of the year, and emissions. How different phases of each oscillation affect aerosol loading is projected using these regressions. AOD responses are identified for each oscillation, with particularly strong responses from El Niño–Southern Oscillation (ENSO) and the Pacific decadal oscillation (PDO). As ENSO can be projected a few months in advance and improvements in linear inverse modeling (LIM) may yield a similar predictability for the PDO, results of this study offer opportunities to improve the predictability of China’s severe wintertime haze events and to inform policy options that could mitigate subsequent health impacts.