Integrated assessment and uncertainty analysis of air pollution emission control and health impacts

The new National Ambient Air Quality Standard for particulate matter will require further reductions in PM concentrations and additional expenditures for most US cities. This thesis develops and applies a methodology for quantifying the health benefits and cost of potential reductions in ambient PM and the contributions of uncertainties to the integrated assessment.; Stochastic distributions and extreme value theory provide an important tool for analyzing air quality data in the context of public policy and regulatory assessment. The violations of air quality standards and the frequency of air pollution episodes can be predicted and used for comparisons with probability based standards. The statistical distributions and the associated uncertainties of predicted PM concentrations of selected control strategies are predicted by a precursor regression model. The methods are illustrated for Los Angeles (LA) County. The results indicate that the statistical model could be subject to bias as a result of omitted factors and/or confounding variables.; Total personal exposure to particulate matter originated from ambient sources is taken into account to re-examine the human health effects derived in existing PM epidemiological studies where only ambient measurements of PM concentrations were used. This study found that predicted risks of premature mortality associated with ambient PM concentrations estimated from the simulated individual exposure model are generally higher, with wider uncertain intervals, than those estimated from the epidemiology model that uses only ambient measurement data.; A decision-making framework is developed to translate emission reductions into local fine particle concentrations and estimate the expected risk of PM related premature mortality and the control costs of emission reductions. Cost and threshold values contribute the most important source of uncertainty to the estimates of the costs per percent annual premature mortality risk reduction.; The integrated assessment model developed in this thesis attempts to offer a simplified overall picture of the processes and consequences that are causally relevant, but by no means comprehensive due to the integration effort. Future research is needed to better assess control costs; to evaluate the degree of nonlinearity in ambient PM concentration response; and to identify the biological existence/mechanism of a threshold for the sensitive population(s) at risk.