University of California Press, Elementa: Science of the Anthropocene, (4), 2016
DOI: 10.12952/journal.elementa.000111
Full text: Download
Abstract The cycling of mercury between ocean and atmosphere is an important part of the global Hg cycle. Here we study the regional contribution of the air-sea exchange in the North- and Baltic Sea region. We use a newly developed coupled regional chemistry transport modeling (CTM) system to determine the flux between atmosphere and ocean based on the meteorological model COSMO-CLM, the ocean-ecosystem model ECOSMO, the atmospheric CTM CMAQ and a newly developed module for mercury partitioning and speciation in the ocean (MECOSMO). The model was evaluated using atmospheric observations of gaseous elemental mercury (GEM), surface concentrations of dissolved gaseous mercury (DGM), and air-sea flux (ASF) calculations based on observations made on seven cruises in the western and central Baltic Sea and three cruises in the North Sea performed between 1991 and 2006. It was shown that the model is in good agreement with observations: DGM (Normalized Mean Bias NMB=-0.27 N=413), ASF (NMB=-0.32, N=413), GEM (NMB=0.07, N=2359). Generally, the model was able to reproduce the seasonal DGM cycle with the best agreement during winter and autumn (NMBWinter=-0.26, NMBSpring=-0.41, NMBSummer=-0.29, NMBAutumn=-0.03). The modelled mercury evasion from the Baltic Sea ranged from 3400 to 4000 kg/a for the simulation period 1994–2007 which is on the lower end of previous estimates. Modelled atmospheric deposition, river inflow and air-sea exchange lead to an annual net Hg accumulation in the Baltic Sea of 500 to 1000 kg/a. For the North Sea the model calculates an annual mercury flux into the atmosphere between 5700 and 6000 kg/a. The mercury flux from the ocean influenced coastal atmospheric mercury concentrations. Running CMAQ coupled with the ocean model lead to better agreement with GEM observations. Directly at the coast GEM concentrations could be increased by up to 10% on annual average and observed peaks could be reproduced much better. At stations 100km downwind the impact was still observable but reduced to 1–3%.