Evaluation of discrepancy between measured and modeled oxidized mercury species
1McGill University, Atmospheric and Oceanic Sciences, 801 Sherbrooke Street West, Montreal, QC, H3A 2K6, Canada
2Independent Researcher, 4998 Maisonneuve West, Westmount, QC, H3Z 1N2, Canada
3Air Quality Research Division, Environment Canada, 2121 Transcanada Highway, Dorval, QC, H9P 1J3, Canada
4Air Quality Research Division, Environment Canada, 4905 Dufferin Street, Toronto ON M3H 5T4, Canada
5McGill University, Atmospheric and Oceanic Sciences and Chemistry, 805 Sherbrooke Street West, Montreal, QC, H3A 2K6, Canada
Abstract. Zhang et al. (2012a), in a recent report, compared model estimates and new observations of oxidised and particulate mercury species (Hg2+ and Hgp) in the Great Lakes region and found that the sum of Hg2+ and Hgp varied between a factor of 2 to 10 between measurements and model. They suggested too high emission inputs and too fast oxidative conversion of Hg0 to Hg2+ and Hgp, as possible causes. This study quantitatively explores in detail the uncertainties in measurements, in addition to the above concerns and speciation of mercury near emission sources in the model to better understand these discrepancies in the context of oxidized mercury, i.e. gaseous (Hg2+) and particulate (Hgp) mercury. These include sampling efficiency, composition of sample, interfering species and calibration errors for measurements and in-plume reduction processes. Sensitivity simulations using Global/Regional Atmospheric Heavy Metals Model (GRAHM) were performed to analyze the role of in-plume reduction on ambient concentrations and deposition of mercury in North America. The discrepancy between simulated and observed concentrations of Hg2+ and Hgp was found to be reduced when a ratio for Hg0:Hg2+:Hgp in the emissions was changed from 50:40:10 (as specified in the original inventories) to 90:8:2 to account for in-plume reduction of Hg0 processes. A significant reduction of the root mean square error (e.g., 19.22 to 11.3 pg m−3 for New Jersey site NJ54) and bias (67.8 to 19.3 pg m−3 for NJ54) for sampling sites in the Eastern United States and Canada, especially for sites near emission sources was found. Significant improvements in the spatial distribution of wet deposition of mercury in North America was noticed. Particularly, over-prediction of wet deposition near anthropogenic sources of mercury was reduced by 43%. On a regional scale, estimated wet deposition improved by a factor of 2 for areas with more than 12 μg m−2 yearly average wet deposition. Model sensitivity simulations show that the measured concentration of oxidized mercury is too low to be consistent with measured wet deposition fluxes in North America. This improvement by a factor of 2 and measurement uncertainties within a factor of 3 to 8 provides a reasonable rationale for the discrepancy of a factor of 2–10 determined by Zhang et al. (2012a).