S3 (II) The North American mercury speciation networks: Analysis and modeling results

Friday, 29 July, 2011

FS3-O9 — 11:00-11:15
Author: LIN, C. Jerry1
(1)Lamar University, Jerry.Lin@lamar.edu

Elemental mercury (Hgo) is a long-lived air pollutant subject to transboundary transport. The long-range transport followed by oxidation and deposition causes concerns of mercury contamination in the receptor regions. There have been a number of atmospheric modeling studies that apportion the contribution to mercury deposition from emissions in various source regions at global and local scales. However, none of the source apportionment work was performed based on emission source sectors at a regional scale, which is important for the formulation of emission reduction strategies.

In this study, the mercury model of the USEPA Community Multi-scale Air Quality modeling system (CMAQ-Hg v4.6) was applied to determine the source contribution to mercury deposition from emissions by electric generating units (EGU), iron and steel industry (IRST), industrial point sources excluding EGU and IRST (OIPM), the rest of anthropogenic sources (RA), natural processes (NAT), and out-of-boundary transport (BC) in the contiguous United States (CONUS). The CONUS region was divided into six sub-regions to understand the regional differences. Annual simulations were performed to quantify the seasonal variability of source apportionment. It was found that, on annual basis, dry deposition accounts for two-thirds of total deposition in the CONUS, mainly contributed by reactive gaseous mercury (ca. 60 % of total deposition). The contribution from large point sources can be as high as 75% near (< 100 km) the emission sources. Out-of-boundary transport contributes from 68% (Northeast region) to 91% (West Central region) of total deposition. Excluding the contribution from out-of boundary transport, EGU contributes to nearly 50% of deposition caused by CONUS emissions in the Northeast, East and East Central regions, while emissions from natural processes are more important in the Pacific and West Central regions (contributing up to 40% of deposition). Deposition is greater in warm seasons due to stronger Hgo oxidation. However, the contribution from anthropogenic sources becomes smaller in warm seasons because of stronger vertical mixing that facilitates transport.

FS3-O10 — 11:15-11:30
Authors: ZHANG, Yanxu1, JAEGLÉ, Lyatt1, HOLMES, Christopher D.2, JACOB, Daniel J.2, VAN DONKELAAR, Aaron3, MARTIN, Randall3
(1) University of Washington, yanxuz@atmos.washington.edu; (2) Harvard University; (3) Dalhousie University;

As a potent neurotoxin, atmospheric mercury (Hg) bioaccumulates in food webs and eventually affects human health once deposited to aquatic environment. Wet deposition of Hg displays high spatial variability, which is difficult to resolve with global scale models. While higher resolution regional models can capture this variability, they suffer from a high sensitivity to assumed lateral conditions provided by global models with different assumptions about emissions, chemistry, deposition, and meteorology. We have developed a nested Hg-oxidant capability for the GEOS-Chem chemical transport model, with a horizontal resolution of 1/2° latitude by 2/3° longitude over N. America. Boundary conditions are provided by a global GEOS-Chem Hg simulation at 4°x5° resolution using the same emissions, chemistry, deposition, and meteorological fields as the high-resolution nested model. We assume that bromine atom is the sole oxidant for Hg. We use bromine concentrations archived from a GEOS-Chem bromine simulation. The nested model improves our simulation of surface total gaseous mercury concentration level compared with the global model. The coefficient of determination for comparison with observations from Canadian Atmospheric mercury Measurement Network (CAMNET) and Atmospheric Mercury Network (AMNET) increases from 25% to 58%. The nested model also shows improved skill at capturing the high spatial and temporal variability of Hg wet deposition over North America at the Mercury Deposition Network (MDN) sites, especially over the Gulf Coast region. This is largely caused by an improved representation of the episodic convective precipitation occurring at small spatial scales by the nested model over this region. In order to reproduce observed wet deposition fluxes and gaseous oxidized mercury concentrations over the Ohio River Valley, we need to include in-plume degradation of oxidized mercury coal fired power plants. We find the contribution of regional anthropogenic sources to the annual Hg wet deposition flux ranges from 13% of the southeast United States to 32% of the Ohio River Valley.

FS3-O11 — 11:30-11:45
Authors: KOS, Gregor1, RYZHKOV, Andrei1, DASTOOR, Ashu1
(1)Environment Canada, gregor.kos@ec.gc.ca

A recent report of concurrent measurements and model runs for oxidised (Hg2+) and particle bound mercury (Hgp) has indicated differences between observations and model simulations (Zhang 2011). Estimations for short-lived Hg2+ species are challenging, because of low concentrations, typically 1-10% of elemental gaseous mercury (Hg0) in the lower pg m-3 range. Measurements and modeling of Hg2+ and Hgp are of importance, because of their reactivity and regional implications via comparatively fast wet/dry deposition and to a lesser extent atmospheric interconversion reactions.

A critical analysis of measurement data is presented including an assessment of instrumental uncertainties. Cold vapour atomic fluorescence (CVAFS) is widely employed for the determination of Hg0. Denuder (for Hg2+) and quartz filter (for Hgp) samplers are used for reactive species followed by CVAFS detection. These species remain operationally defined, while models explicitly specify chemical reactions. Data output and statistics from CVAFS instruments varies considerably. Local effects such as immediate sampling environment and inlet configuration are not considered in a regional and global models, but have significant influence on the collected data.

Uncertainties in model estimates include mercury speciation in anthropogenic emission inventories and post-stack atmospheric processes, e.g., plume chemistry. Rate constants for chemical species and partitioning coefficients are to be considered and so is temporal variability associated with emission inventories. Heterogeneous chemistry losses in the troposphere through reduction on particulates could be of importance. Also, Hg2+ species tend to adsorb on aerosols, as especially HgCl2 is known to be very “sticky” playing an important role for Hg2+ removal from the atmosphere.

A quantitative analysis of uncertainties using continental scale model data from the Global/Regional Atmospheric Heavy Metals Model (GRAHM) and measured data from the Atmospheric Mercury Network (AMNet) and selected Canadian sites is presented.

L. Zhang, P. Blanchard, D. Johnson, A. Dastoor, A. Ryzhkov, C.J. Lin, K. Vijayaraghavan, D. Gay, T. Holsen, J. Huang, J. Graydon, V.L. St. Louis, M.S. Castro, E.K. Miller, F. Marsik, J. Lu, L. Poissant, M. Pilote, K.M. Zhang, Analysis of Modeled Mercury Dry Deposition over the Great Lakes Region, Environmental Pollution (2011) submitted

FS3-O12 — 11:45-12:00
Author: MOORE, Christopher1
(1) University of Maryland Center for Environmental Science, cmoore@umces.edu

Surface atmosphere exchange of gaseous elemental mercury (GEM, Hg0) can be important to the biogeochemical cycling of mercury at background soil sites. For this study we examined GEM exchange on a large areal scale (100s of meters) using micrometeorological techniques. GEM fluxes were measured with the modified Bowen ratio technique along with 30 other atmospheric chemistry and meteorological variables from 7/6/2009 to 7/6/2010 in a ridge top field surrounded by deciduous forest at the Atmospheric Mercury Network (AMNet) MD08 site in western Maryland. Mean GEM flux for the year was -0.63 ± 31.0 ng m-2 h-1 and was not significantly different among seasons. Net GEM flux for the year was -3.33 µg m-2 y-1, roughly half of wet mercury deposition. GEM flux was not correlated with any of the 30 atmospheric chemistry and meteorological variables measured. To determine the factors that may be related to GEM fluxes at MD08, we examined emission and deposition events separately. Annual mean GEM emission was 15.3 ± 27.9 ng m-2 h-1 while annual mean GEM deposition was -14.6 ± 26.6 ng m-2 h-1. Emission in spring was significantly lower than fall but was not significantly different from summer or winter. Mean GEM deposition in the spring was significantly lower than the other seasons. UVB was positively correlated with emission in summer (r = 0.47), fall (r = 0.28), and spring (r = 0.31) and negatively correlated with deposition in summer (r = -0.38) and spring (-0.41). Wind speed was positively correlated with emission in summer (r = 0.27) and winter (r = 0.24) and negatively correlated with deposition in summer (r = -0.42), fall (r = -0.29), winter (r = -0.26), and spring (r = -0.28). Ambient air ozone concentrations in summer were positively correlated with emissions (r = 0.31) and negatively correlated with deposition (r = -0.40). These relationships indicate that GEM fluxes at MD08 were closely related to soil surface processes and but only weakly related to measured atmospheric processes. UVB and ozone could affect the conversion between GEM and gaseous oxidized mercury (Hg2+) near the soil surface. Wind and turbulence could transport this mercury to or from the surface. Only by separating the fluxes into emission and deposition were we able to elucidate these processes influencing the overall GEM flux at MD08.

Friday, 29 July, 2011