Multiscale models of atmospheric mercury

This dissertation uses multiple modeling approaches to identify the key reactions and processes controlling atmospheric mercury globally, with particular emphasis on the upper troposphere and marine boundary layer. We first calculate the global mean atmospheric lifetime of elemental mercury (Hg 0 ) against oxidation by atomic bromine (Br) in the troposphere by combining kinetic data for the Hg-Br system with modeled global concentrations of tropospheric Br. We obtain a lifetime of 0.5-1.7 years based on the range of kinetic data, implying that oxidation of Hg 0 by Br is a major, and possibly dominant, global sink for Hg 0 . Halogens in marine air are proposed to be major oxidants of Hg 0 . We construct a box model of the marine boundary layer (MBL) to interpret observations of reactive gaseous mercury (RGM) in terms of its sources and sinks. The morning RGM rise and midday maximum are consistent with oxidation of Hg 0 by Br atoms, requiring <2 ppt BrO in most conditions. Oxidation of Hg 0 by Br accounts for 35-60% of the RGM source in our model MBL, with most of the remainder contributed by oxidation of Hg 0 by ozone (5-20%) and entrainment of RGM-rich air from the free troposphere (25-40%). Oxidation of Hg 0 by Cl is minor (3-7%), and oxidation by OH cannot reproduce the observed RGM diurnal cycle, suggesting that it is unimportant.

Finally, we compare a global atmospheric model for mercury (GEOS-Chem) in which atomic bromine is the sole oxidant of Hg 0 with an alternative model in which OH and ozone are oxidants. Our model also includes updated emissions (8600 Mg a -1 ), improved scavenging by ice and snow, a new parameterization of HG II uptake and deposition through sea-salt aerosol, and a coupled snowpack emission model. Bromine concentrations here derive from atmospheric models constrained by source gas distributions and match satellite BrO columns. We find that the observed spatial distribution and seasonal cycle of total gaseous mercury (TGM) are equally consistent with either the bromine-only mechanism or the OH and 0 3 mechanism, while mercury oxidation in summer subsidence events over Antarctica is more consistent with the bromine mechanism. Wet deposition measurements over the US are also consistent with the bromine mechanism, but require some additional bromine in the free troposphere, perhaps from sea-salt. Aircraft observations in Arctic spring find complete Hg 0 oxidation in the lower stratosphere that the model cannot reproduce, except possibly with oxidation by chlorine species in the winter vortex. Atmospheric reduction, which has no known mechanism, is not required to reproduce any major features of mercury observations.