Fig. 3: Drivers of primary organic aerosol (POA) and secondary organic aerosol (SOA) trends in China.

a, b Emission- and meteorology-driven variations of annual population-weighted POA and SOA in China during the two action-plan periods. Percent values represent the POA or SOA changes relative to 2013. c, d Yearly variations of annual population-weighted POA and SOA under fixed-2013-emission or fixed-2013-meteorology scenarios compared to their base-model simulated changes relative to 2013. e Pearson r values for the correlations between normalized annual means of key meteorological parameters and those of POA and SOA simulated under the fixed-2013-emission scenario. f Feature importance of meteorological inputs to the meteorology-driven interannual variations of POA and SOA resolved by random forest (RF) model analysis. CONC stands for the concentrations of POA and SOA used in RF which are the model outputs under the fixed-2013-emission scenario. PBLH, SWR, T(2 m), U(10 m), and V(10 m) represent planetary boundary layer height, surface shortwave radiation, air temperature at 2 m, the east–west and north–south wind components at 10 m, respectively. The subscripts, 1 and tp, denote the first model layer and tropopause, respectively.