Analyzing the Complexity of Texas Petroleum Refinery Upset Air Emissions through Characterization, Dispersion, and Distributional Analysis with Environmental Justice Implications

A key scientific and public policy question is whether refinery upset emissions have significant impacts on local air quality and public health. Therefore, the goal of the work presented in this dissertation is to understand the complexity of upset air emissions from Texas petroleum refineries that might be useful in indentifying distributional impacts and health risks to surrounding populations with the environmental justice implications as a result of upsets. This work has led to several conclusions from the six main studies.;First, upset emissions are significant in both size and occurrence when compared to routine operation emissions. It is determined that these events are not random, being more likely to occur during the summer, in the morning, and early in the workweek. A regional analysis of Port Arthur suggests that upset emissions from co-located refineries are equivalent to having an additional refinery within the region. Second, upset emission estimates are sources of uncertainty and lack uniformity exists not only within a single facility, but also across facilities. Third, although there is no consensus within environmental justice research on the most appropriate method or metric for measuring environmental injustice, we examine three levels of spatial analysis for assessing the impact of air emissions from refineries on environmental equity. Our comparison of these three levels identifies large differences in potentially exposed population demographics with each successive level providing more insights critical to understanding environmental justice implications necessary for sound policy decisions. Fourth, in analyzing 20 years of simulated regional SO2 upset events from Port Arthur, we conclude that upsets do matter when thinking about a population's exposure to pollutants and the additional geographic spread of SO2 concentrations. More specifically, we found that 1) upset events are major contributor to regional air quality increasing of SO2 concentrations between 8,000 and 53,000 ppb and directly causing on average two violations for the 24-hour NAAQS annually and for the 1-hr NAAQS up to 200 grid cells in violation, 2) the amount and geographical reach of upsets events varies greatly from year to year, and 3) environmental justice issues are complicated by the non-uniform timing of upset events and the variable wind patterns of the coastal community. Finally, although upsets are violations of the CAA, internal TCEQ documents indicate that permitted upset emissions can be predicted, which implies that modifying the timing and emission rates of upsets events to minimize concentrations may be practical. Changing two event parameters (start time with 2-hours delay/early and emission rate 20% increase/decrease) suggest that large reductions in population exposure to high SO2 concentrations can be achieved.