Hg 0 + I 2 → HgI 2 Absolute N 2 , 1 atm 296 ± 1 < (1.27 ± 0.58) × 10 -19 Raofie et al. 2 M06-2X/aug-cc-pVTZ-PP High pressure limit 3.94 × 10 -14 T 1.06 e -159080/RT Auzmendi-Murua et al. 3 Hg 0 + I → HgI RRKM/B3LYP N 2 , 1 atm 180-400 4.0 × 10 -13 (T/298) -2.38 Goodsite et al. 4 Hg 0 + Br 2 → HgBr 2 Absolute Air, N 2 , 1 atm 298 ± 1 < (9 ± 2) × 10 -17 Ariya et al. 5 Absolute Air, 1 atm ∼298 No reaction detected Sumner et al. 6 Absolute Air, 1 atm 296 (6.0 ± 0.5) × 10 -17 Liu et al. 7 CCSD(T)/AVTZ 1 atm 298-2000 1.62 -9 e -110800/RT Wilcox and Okano 8 M06-2X/aug-cc-pVTZ-PP High pressure limit 4.70 × 10 -14 T 1.06 e -169190/RT Auzmendi-Murua et al. 3 Hg 0 + BrO → HgBrO Relative N 2 , 1 atm 298 10 -15 < k < 10 -13 Raofie and Ariya 9 Hg 0 + Br → HgBr Ab initio N/A, 1 atm 1.01 × 10 -12 e 1738/RT Khalizov et al. 10 RRKM/B3LYP N 2 , 1 atm 200-300 3.7 × 10 -13 (T/298) -2.76 Goodsite et al. 4 ; Goodsite et al. 11 Absolute N 2 , 0.26-0.79 atm 243-293 (1.46 ± 0.36) × 10 -32 [cm 6 molec -2 s -1 ] Donohoue et al. 12 (T/298) (-1.86±1.49) CCSD(T) Ar, 1 atm 260 1.2 × 10 -12 Shepler et al. 13 Relative Air, N 2 , 1 atm 298 ± 1 (3.2 ± 0.9) × 10 -12 Ariya et al. 5 Absolute CF 3 Br, 0.26 atm 397 ~3 × 10 -16 molec -1 s -1 Greig, G. et al. 14 CCSD(T)/AVTZ 1 atm 298-2000 6.64 × 10 -14 (T/298) -0.859 Wilcox and Okano HgBr + Br → HgBr 2 Absolute CF 3 Br, 0.26 atm 397 ~7 × 10 -14 Greig, G. et al. 14 RRKM/B3LYP N 2 , 1 atm 180-400 2.5 × 10 -10 (T/298) -0.57 Goodsite et al. 4 CCSD(T)/AVTZ 1 atm 298-2000 3.32 × 10 -12 (T/298) -9.18 Wilcox and Okano CCSD(T)/aVTZ 1 atm 298 6.33 × 10 -11 Dibble et al. 15 ; Wang et al.

Abstract. Sun-lit snow is increasingly recognized as a chemical reactor that plays an active role in uptake, transformation, and release of atmospheric trace gases. Snow is known to influence boundary layer air on a local scale, and given the large global surface coverage of snow may also be significant on regional and global scales. We present a new detailed one-dimensional snow chemistry module that has been coupled to the 1-D atmospheric boundary layer model MISTRA. The new 1-D snow module, which is dynamically coupled to the overlaying atmospheric model, includes heat transport in the snowpack, molecular diffusion, and wind pumping of gases in the interstitial air. The model includes gas phase chemical reactions both in the interstitial air and the atmosphere. Heterogeneous and multiphase chemistry on atmospheric aerosol is considered explicitly. The chemical interaction of interstitial air with snow grains is simulated assuming chemistry in a liquid-like layer (LLL) on the grain surface. The coupled model, referred to as MISTRA-SNOW, was used to investigate snow as the source of nitrogen oxides (NOx) and gas phase reactive bromine in the atmospheric boundary layer in the remote snow covered Arctic (over the Greenland ice sheet) as well as to investigate the link between halogen cycling and ozone depletion that has been observed in interstitial air. The model is validated using data taken 10 June–13 June, 2008 as part of the Greenland Summit Halogen-HOx experiment (GSHOX). The model predicts that reactions involving bromide and nitrate impurities in the surface snow can sustain atmospheric NO and BrO mixing ratios measured at Summit, Greenland during this period.

Episodes of high bromine levels and surface ozone depletion in the springtime Arctic are simulated by an online air-quality model, GEM-AQ, with gas-phase and heterogeneous reactions of inorganic bromine species and a simple scheme of air-snowpack chemical interactions implemented for this study. Snowpack on sea ice is assumed to be the only source of bromine to the atmosphere and to be capable of converting relatively stable bromine species to photolabile Br₂ via air-snowpack interactions. A set of sensitivity model runs are performed for April 2001 at a horizontal resolution of approximately 100 km&times;100 km in the Arctic, to provide insights into the effects of temperature and the age (first-year, FY, versus multi-year, MY) of sea ice on the release of reactive bromine to the atmosphere. The model simulations capture much of the temporal variations in surface ozone mixing ratios as observed at stations in the high Arctic and the synoptic-scale evolution of areas with enhanced BrO column amount ("BrO clouds") as estimated from satellite observations. The simulated "BrO clouds" are in modestly better agreement with the satellite measurements when the FY sea ice is assumed to be more efficient at releasing reactive bromine to the atmosphere than on the MY sea ice. Surface ozone data from coastal stations used in this study are not sufficient to evaluate unambiguously the difference between the FY sea ice and the MY sea ice as a source of bromine. The results strongly suggest that reactive bromine is released ubiquitously from the snow on the sea ice during the Arctic spring while the timing and location of the bromine release are largely controlled by meteorological factors. It appears that a rapid advection and an enhanced turbulent diffusion associated with strong boundary-layer winds drive transport and dispersion of ozone to the near-surface air over the sea ice, increasing the oxidation rate of bromide (Br^&minus;) in the surface snow. Also, if indeed the surface snowpack does supply most of the reactive bromine in the Arctic boundary layer, it appears to be capable of releasing reactive bromine at temperatures as high as −10 °C, particularly on the sea ice in the central and eastern Arctic Ocean. Dynamically-induced BrO column variability in the lowermost stratosphere appears to interfere with the use of satellite BrO column measurements for interpreting BrO variability in the lower troposphere but probably not to the extent of totally obscuring "BrO clouds" that originate from the surface snow/ice source of bromine in the high Arctic. A budget analysis of the simulated air-surface exchange of bromine compounds suggests that a "bromine explosion" occurs in the interstitial air of the snowpack and/or is accelerated by heterogeneous reactions on the surface of wind-blown snow in ambient air, both of which are not represented explicitly in our simple model but could have been approximated by a parameter adjustment for the yield of Br₂...