Estimating the effect of past emissions controls on ozone levels in Georgia: Identifying a “safety buffer” for ensuring long-term attainment

Both Atlanta and Macon, Georgia were recently designated by the EPA as non-attainment areas for the 8-hour ozone standard. The Georgia Environmental Protection Division (GA EPD) is implementing a state of the science modeling system to consider various emissions control strategies for attaining the ozone standard. Under the EPA's attainment demonstration methodology, photochemical models are applied in a relative sense to predict the ratio between future and base year ozone levels; that ratio is then multiplied by base period observed “design value” (DV) pollutant concentration to determine attainment. However, given the sensitivity of ozone to meteorological conditions and the significant variations observed in the forth-highest value (FHV) per season (used to calculate the DV, which is an average of the FHV for the last three years), this method may fail to ensure future attainment if the base period was unusually cool or humid.

For these reasons, the GA EPD is supplementing its modeling efforts with observational analyses to assess the variability in local ozone levels and the relative extent to which recent declines in ozone DVs can be attributed to emissions trends and meteorological variability. Analyzing trends in emissions inventories and using various emissions models, we have identified that since 1998 emissions of ozone precursors in Georgia have dropped from both the power and transportation sectors. Based on the method of Cox and Chu (1993), we used a Weibull distribution fit to calculate daily maximum 8-hour ozone concentrations based on a suite of meteorological parameters (short-term effect) and using calendar year as a surrogate for emission levels (long-term trend) for the period 1998-2004. During this period, the ozone FHV at a central Atlanta monitor declined by more than 30 ppb. Results indicate that ozone FHVs at Atlanta area monitors have dropped by approximately 2-3% a year due to changes in emissions alone. This signifies that if the ozone-conducive meteorology of 1998-2000 were to recur with present emissions levels, the actual FHVs would be lower by at least 10 ppb at most sites compared to what had been experienced originally. The obtained Weibull fit was also used to estimate the likelihood of extreme ozone conducive conditions using a dataset of more than 40 years of meteorology. These results are being used to estimate a “safety buffer” for ensuring long-term attainment, beyond the EPA required attainment demonstration for a specific year.