Search

Exposure to ambient fine particulate matter (PM2.5) is a leading risk factor for the global burden of disease. However, uncertainty remains about PM2.5sources. We use a global chemical transport model (GEOS-Chem) simulation for 2014, constrained by satellite-based estimates of PM2.5to interpret globally dispersed PM2.5mass and composition measurements from the ground-based surface particulate matter network (SPARTAN). Measured site mean PM2.5composition varies substantially for secondary inorganic aerosols (2.4–19.7 μg/m3), mineral dust (1.9–14.7 μg/m3), residual/organic matter (2.1–40.2 μg/m3), and black carbon (1.0–7.3 μg/m3). Interpretation of these measurements with the GEOS-Chem model yields insight into sources affecting each site. Globally, combustion sectors such as residential energy use (7.9 μg/m3), industry (6.5 μg/m3), and power generation (5.6 μg/m3) are leading sources of outdoor global population-weighted PM2.5concentrations. Global population-weighted organic mass is driven by the residential energy sector (64%) whereas population-weighted secondary inorganic concentrations arise primarily from industry (33%) and power generation (32%). Simulation-measurement biases for ammonium nitrate and dust identify uncertainty in agricultural and crustal sources. Interpretation of initial PM2.5mass and composition measurements from SPARTAN with the GEOS-Chem model constrained by satellite-based PM2.5provides insight into sources and processes that influence the global spatial variation in PM2.5composition.