AN ATMOSPHERIC DISPERSION MODEL FOR THE SUDBURY, ONTARIO, AREA

A mathematical model was developed and tested to predict the relationship between sulphur oxide and trace metal emissions from smelters in the Sudbury, Ontario area, and atmospheric, precipitation, lake water and sediment chemistry.;Important improvements over existing models include (1) near- and far-field conditions treated in a single model; (2) direct linkage of crosswind dispersion to hourly meteorological observations; (3) utilization of maximum to minimum range of input parameters to realistically model the range of outputs; (4) direct linkage of the atmospheric model to a lake model.;Precipitation chemistry as calculated by the atmospheric model is related to lake water and sediment chemistry utilizing a mass balance approach and assuming a continuously stirred reactor (CSTR) model to describe lake circulation. All inputs are atmospheric, modified by hydrology, soil chemistry and sedimentation.;Model results were tested by comparison with existing atmospheric and precipitation chemistry measurements, supplemented with analyses of lake water and sediment chemistry collected in a field program.;The model consists of atmospheric and lake chemistry portions. The atmospheric model is a Gaussian crosswind concentration distribution modification to a box model with a uniform vertical concentration gradient limited by a mixing height. In the near-field Briggs' plume rise and vertical dispersion terms are utilized. Oxidation, wet and dry deposition mechanisms are included to account for the gas, liquid and solid phases separately.;Eight pollutant species were selected for modeling: sulphur dioxide, sulphate ion, hydrogen ion, copper, nickel, lead, zinc, and iron.;The model effectively predicts precipitation chemistry within 150 km of Sudbury, with an average prediction to measurement ratio of 90 percent. Atmospheric concentrations are effectively predicted within 80 km, with an average prediction to measurement ratio of 81 percent.;Lake chemistry predictions are good, with an average prediction to measurement ratio of 140 percent, however, the pattern of results indicates a more complex lake model is required to account for limnological and watershed factors.;Sediment chemistry predictions are good, with an average prediction to measurement ratio of 166 percent.