Please use this identifier to cite or link to this item:
http://hdl.handle.net/11375/6319
Title: | An Atmospheric Dispersion Model For The Sudbury, Ontario, Area |
Authors: | Huhn, Frank J. |
Advisor: | Kramer, J. R. |
Department: | Geology |
Keywords: | Geology;Geology |
Publication Date: | Mar-1981 |
Abstract: | <p>The objectives of the study are to predict the relationship of atmospheric emissions from smelters to atmospheric and precipitation chemistry, and to lake water and sediment chemistry.</p> <p>A mathematical model was developed and tested to simulate atmospheric, precipitation, lake water and sediment chemistry of selected lakes in an approximate 0.3 x 10⁶ square kilometre area around the city of Sudbury, Ontario.</p> <p>The model consists of two basic parts; an atmospheric portion and a lake chemistry portion. The atmospheric model is a Gaussian crosswind distribution modification to a box model with a uniform vertical concentration gradient limited by a mixing height (for the far field). In the near field, plume rise and vertical dispersion terms are utilized. First order chemical reaction kinetics are utilized to model the production of hydrogen ion and sulphate ion by the oxidation of sulphur dioxide. Oxidation processes in both the gas phase and the liquid phase are incorporated.</p> <p>Losses en route are accounted for by the incorporation of deposition mechanisms. Losses result from the operation of these mechanisms over a period determined by plume travel time from the source to the receptor. Separate wet and dry deposition mechanisms acting on the different physical phases of pollutant species in the atmosphere are considered. These include pollutants in the gas phase, the liquid phase (dissolved in cloud droplets) and particulates or aerosols.</p> <p>The mass flux of pollutant onto a unit surface (loading rate) at the receptor is accounted for by the incorporation of deposition mechanisms. While the physical processes involved are the same as those producing enroute losses, loadings result from the operation of these mechanisms at the receptor over the time period of model operation.</p> <p>The continued oxidation of sulphur dioxide in samplers after deposition is also modeled.</p> <p>The model can be operated in the Gaussian crosswind mode or the box mode. Advection, dispersion and boundary conditions are described by using hourly and daily meteorological data from a series of meteorological stations in the study area. These data are preprocessed by a subprogram. Provision for six alternative schemes for combining the data from the meteorological stations are included.</p> <p>Important improvements over previous models include</p> <p>-computations for near and far field conditions in a single model</p> <p>-direct linkage of crosswind dispersion to daily and hourly meteorological parameters.</p> <p>-utilization of minimum and maximum input driving parameters and theoretical parameter measurement accuracy to realistically model the range of outputs</p> <p>-direct linkage of the atmospheric model to a lake water and sediment model.</p> <p>The second portion of the model is the 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. In the model surface water, the epilimnion; and the hypolimnion are considered as a single unit.</p> <p>All inputs to the lake model are determined by predicted atmospheric and precipitation chemistry, modified by hydrology, soil chemistry and sedimentation in the drainage basin. Lake model outputs are stream flow and sedimentation.</p> <p>A field program was carried out to collect sediment and lake chemistry data with which to test the lake model. These observations of lake water and sediment chemistry, geological observations and measurements supplement atmospheric chemistry, precipitation chemistry, meteorological and topographic data and available literature concerning the Sudbury study area.</p> <p>The study area chosen surrounds the smelters at Sudbury and Falconbridge. Eight pollutant species were selected for modeling--sulphur dioxide, sulphate ion, hydrogen ion, copper, nickel, lead, zinc and iron particulates. The INCO and Falconbridge copper/nickel ore smelters in the Sudbury, Ontario area are among the largest single sources of sulphur dioxide in the world, emitting approximately 1 percent of the anthropogenic total. Significant quantities of particulates containing iron, copper, nickel, lead and other trace elements are also emitted. Sulphur dioxide, copper, iron and nickel are important smelter emissions. Zinc and iron are also indicators of natural geological processes affecting atmospheric chemistry, while lead and zinc also provide indicators of urban processes other than smelter operation.</p> <p>The Sudbury source is relatively isolated from other sources of comparable magnitude, and topographic and meteorological conditions are relatively uniform across the study area. This presents advantages for model verification.</p> <p>The study area contains large numbers of accessible lakes in a variety of geological formations, including small unbuffered lakes in small bedrock basins in the Lorraine Quartzite formation. Several of these lakes were examined in comparison to large basin lakes with higher buffering capacities in order to effectively gauge the effects of atmospheric inputs from Sudbury on lakewater and sediment chemistry.</p> <p>Results indicate that the model effectively predicts precipitation chemistry within approximately 150 kilometers of Sudbury. Predictions of precipitation chemistry range from approximately 22±36 percent of measured values for hydrogen ion to approximately 242±179 for copper. The average ratio of predicted to measured precipitation chemistry values is approximately 90±80 percent. Beyond 150 kilometres the effects of other pollutant sources begin to dominate. Utilization of Sudbury centered sources only in the model is invalid at these distances.</p> <p>Predictions of precipitation chemistry are calculated from model predictions of atmospheric concentrations. Therefore the accuracy of the atmospheric concentration predictions defines an upper limit to the accuracy of the precipitation chemistry predictions. Therefore atmospheric concentration predictions will be at least as accurate as precipitation chemistry predictions. Results indicated that excellent prediction of atmospheric concentrations are achieved with the model within 80 kilometers of the smelters. Predictions of atmospheric concentrations range from approximately 69±26 percent of measured values for sulphur dioxide, to approximately 105±37 percent of measured values for nickel. The average ratio of predicted to measured atmospheric concentrations is approximately 81±21 percent.</p> <p>Predictions of lake water chemistry are calculated from predictions of precipitation chemistry, hence the accuracy of precipitation chemistry predictions defines an upper limit to the accuracy of lake chemistry predictions. Results indicate that good predictions of lake chemistry are achieved with the model. Predictions of lake water chemistry range from approximately 10±6 percent of measured values for hydrogen ion, to approximately 270±265 percent of measured values for zinc. The average ratio of predicted to measured lake chemistry values is approximately 140±124 percent. Ratios of predicted to measured lake water chemistry exhibit a clear bimodal pattern for pH, copper, nickel, zinc and iron. pH predictions for large drainage basin lakes are high, and pH predictions for small drainage basin lakes are low. Geological controls of lake chemistry, especially of pH and the control pH exerts on trace metal chemistry, may account for this pattern. These results indicate that adequate lake chemistry modeling requires a more complex model then a simple CSTR mass balance.</p> <p>Predictions of sediment chemistry are calculated from model predictions of lake water chemistry. Therefore the accuracy of sediment chemistry predictions are limited by the accuracy of lake water chemistry predictions. Results indicate that good predictions are achieved with the model. Predictions of sediment chemistry range from approximately 102±78 percent of measured values for nickel, to 263±228 percent of measured values for lead. The average ratio of predicted to measured sediment chemistry is approximately 166±122 percent.</p> |
URI: | http://hdl.handle.net/11375/6319 |
Identifier: | opendissertations/1638 2055 1240340 |
Appears in Collections: | Open Access Dissertations and Theses |
Files in This Item:
File | Size | Format | |
---|---|---|---|
fulltext.pdf | 11.24 MB | Adobe PDF | View/Open |
Items in MacSphere are protected by copyright, with all rights reserved, unless otherwise indicated.