Aircraft-Based Measurements and Modeling of Submicron Aerosols over Northeastern U.S. during the WINTER 2015 Campaign

Unhealthy levels of submicron aerosols (PM₁) are the main cause of poor air quality in urban areas in the northeastern U.S. during winter. These particles are either emitted directly (primary) or formed in situ from oxidation of precursor gases (secondary). The concentrations of secondary particles depend on temperature, relative humidity, precursor emissions, oxidation capacity, concentrations of gas-phase counterparts, and the concentration and composition of existing particles. Few studies have examined the interplay of these processes during winter. Here, we use aircraft-based observations collected during the recent Wintertime INvestigation of Transport, Emissions and Reactivity (WINTER) campaign (Feb-March 2015), together with the GEOS-Chem chemical transport model, to investigate the sources and chemical processes governing wintertime PM₁ over the northeastern U.S. We focus on sulfate, nitrate, ammonium, and organic aerosol (OA) particles. The mean observed concentration of PM₁ between the surface and 1 km was 4 μg m -3 , about 30% of which was composed of sulfate, 15% each of ammonium and nitrate, and the remaining 40% of OA. The mean pH of the observed aerosols is calculated to be 0.7. The model reproduces observed concentrations of SO₂ and sulfate, suggesting that anthropogenic emissions are well represented by the EPA NEI 2011 inventory and that oxidation chemistry of SO₂ is well captured. In contrast, we find that NH₃ and ammonium concentrations are overestimated by a factor of 2 in the NEI 2011 inventory, which could be due to NH₃ emission suppression during the unusually cold winter of 2015. Similar to previous studies, we find that the GEOS-Chem model significantly overestimates HNO₃ and aerosol nitrate. We use WINTER observations to constrain the heterogeneous production of HNO₃ via N₂O₅ hydrolysis, as well as loss of HNO₃ via dry deposition and uptake on sea-salt particles downwind of the eastern seaboard. The dependence of nitrate gas-particle partitioning on particle pH in the model is consistent with the observations. Despite higher HNO₃ concentrations, the mean particle pH in the model is too high (1.6), likely due to counteracting effect of higher NH₃ concentrations on particle pH. The modeled OA concentrations are in reasonable agreement with the observations, but the proportion of the secondary component is greatly underestimated (by a factor of 10 or more) in comparison to the oxidized organic aerosol (OOA) fraction obtained from factor analysis of the aerosol mass spectrometer (AMS) measurements. We test commonly used parameterizations for modeling the gas-particle partitioning and aging of organic aerosols against these AMS measurements. Additionally, we present the contribution of various sources to the observed PM₁ during the campaign.