Environmental Research Letters (Jan 2013)
Changes in the frequency and return level of high ozone pollution events over the eastern United States following emission controls
Abstract
In order to quantify the impact of recent efforts to abate surface ozone (O _3 ) pollution, we analyze changes in the frequency and return level of summertime (JJA) high surface O _3 events over the eastern United States (US) from 1988–1998 to 1999–2009. We apply methods from extreme value theory (EVT) to maximum daily 8-hour average ozone (MDA8 O _3 ) observed by the Clean Air Status and Trends Network (CASTNet) and define O _3 extremes as days on which MDA8 O _3 exceeds a threshold of 75 ppb (MDA8 O _3 >75). Over the eastern US, we find that the number of summer days with MDA8 O _3 >75 declined on average by about a factor of two from 1988–1998 to 1999–2009. The applied generalized Pareto distribution (GPD) fits the high tail of MDA8 O _3 much better than a Gaussian distribution and enables the derivation of probabilistic return levels (describing the probability of exceeding a value x within a time window T) for high O _3 pollution events. This new approach confirms the significant decline in both frequency and magnitude of high O _3 pollution events over the eastern US during recent years reported in prior studies. Our analysis of 1-yr and 5-yr return levels at each station demonstrates the strong impact of changes in air quality regulations and subsequent control measures (e.g., the ‘NO _x SIP Call’), as the 5-yr return levels of the period 1999–2009 correspond roughly to the 1-yr return levels of the earlier time period (1988–1998). Regionally, the return levels dropped between 1988–1998 and 1999–2009 by about 8 ppb in the Mid-Atlantic (MA) and Great Lakes (GL) regions, while the strongest decline, about 13 ppb, is observed in the Northeast (NE) region. Nearly all stations (21 out of 23) have 1-yr return levels well below 100 ppb and 5-yr return levels well below 110 ppb in 1999–2009. Decreases in eastern US O _3 pollution are largest after full implementation of the nitrogen oxide (NO _x ) reductions under the ‘NO _x SIP Call’. We conclude that the application of EVT methods provides a useful approach for quantifying return levels of high O _3 pollution in probabilistic terms, which may help to guide long-term air quality planning.
Keywords