Investigating the Sources of Anthropogenic Wintertime Pollutants in the northeast United States with a Lagrangian Dispersion Model

Abstract

Anthropogenic emissions exhibit different seasonality depending on the source type and species. To investigate the sources of wintertime pollutants over the northeastern United States (NE US), the WINTER (Wintertime INvestigation of Transport, Emission, and Reactivity) aircraft campaign occurred in February-March 2015 and sampled air downwind the North East corridor and over the Ohio River Valley. I used the FLEXible PARTicle (FLEXPART) Lagrangian transport model to investigate the origin of five of the observed pollutants: CO, NOx, SO2, NOy, and HCHO. I examined their geographic location, source type, and the transport time since emission. I found that FLEXPART with the Environmental Protection Agency’s (EPA) 2011 National Emissions Inventory (NEI 2011) leads to a reasonable representation of the observed enhancements of NOx and NOy above background, with a normalized mean bias (NMB) of 2% and 5%, respectively. For CO and SO2, the NMB was larger at -14% and 48% and both had a poor correlation of 0.31 and 0.32. For CO and NOx, I find that 70% of the observed CO enhancement and 46% of the observed NOx enhancement originate from mobile sources. For SO2, 71% of observed enhancement was from the energy generation sector, mostly coal combustion to produce electricity. The top 3 cities contributing to CO and NOx enhancements were: New York (38%, 45%), Washington DC & Baltimore (8.9%,7.4%), Philadelphia (6.6%, 7.8%). For SO2: New York (18%), Pittsburgh (12%), Louisville (12%). About half of the CO enhancement observed was determined to have been emitted less than 24 hours prior to sampling, while 81% of the NOx observed was emitted less than 24 hours earlier and 46% less than 12 hours earlier. For SO2, 88% of the SO2 observed was less that 48 hours old; 34% of that was less than 12 hours old. FLEXPART was used to investigate if primary HCHO or secondary HCHO could account for the large difference between the observations and the emissions as represented in the NEI 2011 inventory. Using a simplified VOC reaction with OH to produce SOA and HCHO, secondary HCHO contribution was determined. The correlation coefficient of predicted HCHO to the observations increased for secondary (0.59) vs primary (0.31). A least-squares linear fit was used to match the magnitude of observed HCHO with a combination of primary and secondary sources. The average percentage of HCHO from primary sources was determined to be 21.5% while 78.5% was from secondary production indicating secondary sources of HCHO are important factor in the winter.