Oxidation of elemental mercury in the atmosphere; Constraints imposed by global scale modelling

We use the global mercury model published by Bergan et al. (1999) to evalua te the potential role of ozone and the hydroxyl radical as gas phase oxidan ts for the oxidation of elemental mercury in the atmosphere. The magnitude of natural and man-made mercury emissions are taken from recent literature estimates. We consider only two mercury reservoirs, elemental mercury, Hg-0 , and the more soluble divalent form, HgII. Wet and dry deposition of HgII is explicitly treated. Applying monthly mean fields of ozone for the oxidation of gas phase Hg-0 a nd using the reaction rate by Hall (1995) yields a global transformation of Hg-0 to HgII which is too slow to keep the simulated concentration of Hg-0 near observed values. This shows that there must be additional important r emoval processes for Hg-0 or that the reaction rate proposed by Hall (1995) is too slow. A simulation in which the oxidation rate was artificially inc reased, so that the global turn-over time of Hg-0 is one year and the simul ated average concentration of Hg-0 realistic, produces latitudinal and seas onal variations in Hg-0 that do not support the hypothesis that gas phase r eaction with O-3 is the major oxidation process for Hg-0. Recent studies indicate that OH may be an important gas phase oxidant for H g-0 (Sommar et al., 2001). Using OH as the sole oxidant and applying the ox idation rate by Sommar et al., we calculate a concentration of Hg-0 well be low (about a factor of three) the observations. By prescribing a slower rat e, corresponding to a turn-over time of Hg-0 of one year, we calculate conc entrations of both Hg-0 in surface air and HgII in precipitation which corr espond reasonably well, both in magnitude and temporal variation, with seas onal observations in Europe and North America. This result supports the sug gestion that the oxidation by OH is an important pathway for the removal of Hg-0. In view of the uncertainties associated with our calculations, this conclusion should still be regarded as tentative.