Distinguishing Emission-Associated Ambient Air PM2.5 Concentrations and Meteorological Factor-Induced Fluctuations.

Although PM2.5 (particulate matter with aerodynamic diameters less than 2.5 μm) in the air originates from emissions, its concentrations are often affected by confounding meteorological effects. Therefore, direct comparisons of PM2.5 concentrations made across two periods, which are commonly used by environmental protection administrations to measure the effectiveness of mitigation efforts, can be misleading. Here, we developed a two-step method to distinguish the significance of emissions and meteorological factors and assess the effectiveness of emission mitigation efforts. We modeled ambient PM2.5 concentrations from 1980 to 2014 based on three conditional scenarios: realistic conditions, fixed emissions, and fixed meteorology. The differences found between the model outputs were analyzed to quantify the relative contributions of emissions and meteorological factors. Emission-related gridded PM2.5 concentrations excluding the meteorological effects were predicted using multivariate regression models, whereas meteorological confounding effects on PM2.5 fluctuations were characterized by probabilistic functions. When the regression models and probabilistic functions were combined, fluctuations in the PM2.5 concentrations induced by emissions and meteorological factors were quantified for all model grid cells and regions. The method was then applied to assess the historical and future trends of PM2.5 concentrations and potential fluctuations on global, national, and city scales. The proposed method may thus be used to assess the effectiveness of mitigation actions.