Air-sea exchange and marine boundary layer atmospheric transformation of hg and their importance in the global mercury cycle

Abstract

The atmosphere is the most important pathway for the worldwide dispersion and transport of Hg (Fitzgerald et al., 1998; Mason et al., 1994). Understanding the global transport and atmospheric transformations of Hg is important because of the ability of Hg deposited to aquatic systems, to be converted to methylmercury (MeHg) and to bioaccumulate through all levels of the food chain. Most of the Hg in the atmosphere is elemental Hg (Hg), which is relatively unreactive with the net average atmospheric residence time of around one year. In addition to Hg, two other atmospheric Hg fractions have been operationally defined based on physicochemical properties-the gaseous ionic Hg11 fraction, which has been termed reactive gaseous Hg (RGHg), and particulate-bound Hg (Hg-P). The speciation of RGHg is not known in detail but based on laboratory studies and the methods of its collection (Landis et al., 2002; Sheu and Mason, 2001; Lindberg and Stratton, 1998; Ariya et al., 2002; Sheu and Mason, 2004), it is assumed to consist of gaseous neutral Hg 11 complexes such as HgCl2, HgBr2, and HgOBr (Balabanov and Peterson, 2003; Kalizov et al., 2003). Such compounds are highly surface-reactive and substantially more water-soluble than Hg. Estimates of dry deposition velocities for RGHg for the open ocean range from 0.5-4 cm s -1 (Laurier et al., 2003), estimated using the formulation of Shahin et al. (2002) and are much higher than those for Hg-P (0.1-0.5 cm s -1), which being mostly derived from high combustion sources, is associated with the fine particulate fraction (Bullock et al., 1997). In many locations, dry deposition could be as important as wet deposition in terms of being a Hg source to the Earth's surface. Global Hg models have identified wet and dry particle deposition and evasion of dissolved gaseous Hg from the ocean as critical pathways for global Hg cycling (Mason et al., 1994; Hudson et al., 1995; Lamborg et al., 1999; Shia et al., 1999). Natural sources of Hg to the atmosphere are mainly in the form of Hg although emissions of Hg-P also occur (e.g., volcanoes, dust) while anthropogenic sources contribute all forms of Hg to the atmosphere (Ebinghaus et al., 1999). In addition, in situ oxidation of atmospheric Hg in the gas phase could be a source of RGHg as Hg can be oxidized by hydrogen peroxide (H2O2), albeit slowly (Tokos et al., 1998), the nitrate (NO3) radical (Sommar et al., 1997) and other reactive nitrogen intermediates, ozone (O3) (Hall, 1995), and the hydroxyl (OH) radical (Sommar et al., 2001). Given the typical atmospheric Hg concentrations of these oxidants, it is unlikely that these homogeneous reactions dominate the Hg oxidation in general, based on estimates of the rate that is required for its removal via wet and dry deposition. However, the gas phase oxidation of Hg by halogen atoms and molecules (Cl, Br, Br2, Cl2; Lin and Pehkonen, 1999; Sliger et al., 2000; Ariya et al., 2002), and potentially by other halogen compounds (e.g., BrCl, BrO), has been recently demonstrated. The reactions with atomic Cl and Br have the larger rate constants (Ariya et al., 2002) and therefore oxidation by these mechanisms may proceed much faster than oxidation by O3 and OH, in certain locations such as the polar region, the marine boundary layer, and at high altitudes (Laurier et al., 2003; Lindberg et al., 2002; Landis and Stevens, see Chapter-7 in this book). Measurements of Hg depletion events in surface-level Arctic air during the three month period following polar sunrise (Schroeder et al., 1998) provided the first indication of the importance of halogen-mediated reactions in Hg oxidation. The fluctuation of total atmospheric Hg (primarily Hg) strongly resembled the fluctuation of ambient O3 concentrations during the same period (Schroeder et al., 1998), suggesting that both species were removed by similar mechanisms. Many more recent studies (Ebinghaus et al., 2002; Temme et al., 2003; Lindberg et al., 2002) have confirmed that such rapid Hg depletion events commonly occur during polar sunrise, being driven by photochemical reactions, in both the Arctic and Antarctic and that the depletion of Hg coincides with the increase in RGHg concentrations (Lindberg et al., 2002). The destruction of O3 during polar sunrise correlates with the production of reactive halogen species (RXS). It is currently thought that O3 destruction is initiated by Br atoms and BrO, and to a lesser extent, by Cl atoms (McConnell et al., 1992; Barrie et al., 1988; Foster et al., 2001; Figure 1). with the sea-salt particles that deposit to and accumulate in the snow pack during winter being the major source of the precursors of Br and Cl atoms in polar regions, e.g., Br2 and BrCl (Finlayson-Pitts et al., 1990; Foster et al., 2001). Similar reactions could oxidize Hg to RGHg. (figure presentation) The X2 precursors of the RXS (Br, Cl, BrO, ClO), such as Br2, BrCl, and Cl2, are also liberated from sea-salt particles in the marine boundary layer (MBL), and the subsequent O3 destruction, has also been measured and described (Mozurkewich, 1995; Vogt et al., 1996; Knipping et al., 2000; Galbally et al., 2000; Hirokawa et al., 1998) and observed in laboratory studies (Oum et al., 1998; Gabriel et al., 2002; Fickert et al., 1999). Additionally, high concentrations of Cl2 have also been detected in coastal air (Spicer et al., 1998). An important intermediate product of the O3 destruction chain reaction, BrO, was detected in the ambient air over the Dead Sea and showed an inverse correlation with O3 (Hebestreit et al., 1999). BrO is the precursor to HOBr, which directly interacts with salt surfaces to regenerate RXS (von Glasow et al., 2002; Vogt et al., 1996; Figure 1). Dickerson et al. (1999) reported large diurnal variations in O3 concentration in the marine boundary layer over the tropical Indian Ocean and found, in modelling their results, that the temporal trend of, and the magnitude of variation in, O3 concentrations were more compatible with the field data when halogen chemistry and RXS formation was included. The presence of RGHg in the MBL in remote locations, and evidence of a diurnal variation in concentration (e.g., Laurier et al., 2003; Hedgecock et al., 2001; 2003) suggests that RGHg is being produced photochemically by reaction with RXS in conjunction with ozone destruction. Laboratory studies also confirm the potential importance of such reactions (Sheu and Mason, 2004; Ariya et al., layer, or in other regimes where there is the presence of halogen-containing aerosol. Principal oxidation reactions involve reactive halogen species such as Br and Cl. The hydoxyl radical is also directly or indirectly involved in mercury oxidation. In addition to halogen-mediated Hg oxidation at the Earth's surface, there is also accumulating evidence for Hg depletion events in the upper atmosphere. Landis and Steven (see Chapter-7 in this book) discuss results from studies at Mauna Loa (3500m) and from aircraft studies off the coast of Florida (60-3500m). In earlier mass balance models it was assumed that dry deposition was not an important component. This was because, as discussed above, dry particulate Hg deposition is much less than wet deposition for most remote environments. However, scavenging of particulate Hg is likely to be an important contributor to wet deposition. For dry deposition of gaseous Hg, it is known that Hg deposition does not normally occur. The more recent studies that have measured RGHg using accepted protocols have shown that, because of its high deposition velocity, RGHg deposition can rival that of wet deposition in locations where RGHg concentrations are in excess of about 20 pg m-3. The relative importance of wet versus dry deposition will be discussed further in this chapter. At the air/water interface, gas exchange of dissolved gaseous mercury (DGHg) is the main "sink", via evasion, for surface ocean Hg. DGHg consists of elemental Hg (Hg) and dimethylmercury (DMHg) (Kim and Fitzgerald, 1988; Mason et al., 1995; Cossa et al., 1997; Lamborg et al., 1999) with Hg being the dominant form of DGHg in the surface ocean. Earlier studies focused on air-sea exchange for the open ocean, both the Atlantic (Cossa et al., 1997; Mason and Sullivan, 1999; Lamborg et al., 1999) and Pacific Oceans (Kim and Fitzgerald, 1988; Mason and Fitzgerald 1993). More recent studies have focused on coastal regions and the Mediterranean (Gardfeldt et al., 2003; Ferrara et al., 2003; Baeyens and Leermakers, 1998; Mason et al., 1999). Many studies have suggested that the estimated evasion rates for Hg from the ocean substantially exceed the current wet plus particulate dry deposition estimates and riverine inputs, suggesting another potential source for upper ocean Hg. One potential reason for this lack of balance is the relative paucity (both temporally and spatially) of DGHg data for the ocean and thus its potentially unrepresentative nature. Accordingly, it has recently been hypothesized that there is a substantial input of Hg to the ocean via dry deposition of RGHg (Mason and Sheu, 2002).