Environmental models to assess regional impacts to support cleaner technology decisions: Case study of perchloroethylene used for dry cleaning in Los Angeles and Chicago

Cleaner industrial process technologies are effective in achieving pollution prevention, but to promote them decision-makers must understand the health and environmental tradeoffs of different alternatives. This research was conducted to build a link between industrial process technology change and reduction in health and environmental impacts. It demonstrated a method by which the impacts of competing technologies maybe quantitatively assessed with the help of environmental models to support decisions to invest in, implement and enforce the truly cleaner alternative.; The research applied a level III fugacity model in a case study to assess behavior of perchloroethylene (PCE) emitted by the dry cleaning industry into different media in two urban regions and two seasons, and estimated excess cancer risk of the public from exposure to the substance. It compared the deterministic and probabilistic approaches, and examined uncertainty arising from input data uncertainty and variability as well as the analyst's choices and assumptions.; Modeling under region-specific conditions was found to be necessary for cleaning technology assessment, as predicted PCE concentrations and excess cancer risks showed significant regional and seasonal variations in ways that could not be predicted by simple measures such as total emissions of PCE. The deterministic approach was found unreliable in assessing PCE concentrations and related health effects. The probabilistic approach, on the other hand, could provide information about the range and centrality of the concentrations and risks, enabling decision-makers to make informed decisions about competing technologies.; Uncertainty in the model's output was found to come from uncertainty and{09}variability in the input physicochemical and landscape parameters, as well as the analyst's choice of values for such factors as modeling domain size, emission strength and dose-response relationship. A sensitivity analysis found that the choice affected not only the modeling results but also the relative importance of different physicochemical and landscape parameters. It also showed that uncertainty in a few input parameters could explain most of the variability in the modeling results in any given situation. Identifying those parameters using the method demonstrated in this research may lead to increased efficiency of the assessment.