II. CLIMATOLOGY OF OZONE AND OZONE PRECURSORS (C)

The ozone climatology research program within SOS aimed at identifying the long-term weather-related characteristics of the southeastern region of the United States that influence the pollution climate of the southern states. The SOS region includes the states of Alabama, Florida, Georgia, Kentucky, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, and Texas. This region includes the warm and humid Atlantic Coast and Gulf Coastal Plains, the moderate elevation hilly Piedmont regions of all ten states, and the high-elevation, nearly boreal Appalachian Mountain areas within the states of North Carolina and Tennessee and foothill areas of Georgia, Alabama, and Mississippi. Much of this large region has a higher intensity of solar radiation and somewhat lower average wind speeds, and a higher frequency of air stagnation events than other parts of the eastern US and Canada.

It is well known that natural conditions such as solar radiation, temperature, relative humidity, and wind speed affect not only the photochemical processes that lead to ozone and particulate matter formation and accumulation in the atmosphere, but also influence the rate and magnitude of air emissions and dispersion of the chemical precursors of ozone and other oxidants (mainly VOC and NOx), as well as PM2.5 (mainly VOC, NOx, SO2, and NH3).

To the extent that these conditions and their impacts in the SOS region are different from those in other regions, both the chemical climatology and optimal strategies for mitigating the photochemical ozone problem in the SOS region may be different from those in other regions. Thus, evidence generated in ozone climatology studies has critically important implications with respect to the possible value of region-specific rather than uniform across-the-country ozone control strategies.

The major results from SOS ozone climatology studies are summarized in the specific scientific findings and policy implications listed below.

C1. The ozone pollution problems in rural and urban areas within the ten-state SOS region (NC, SC, KY, TN, GA, FL, AL, MI, LA, and TX) are somewhat different in character from that of the mid-Atlantic region (VA, MD, DL, NJ) and even more different in character from that of the midwestern (OH, IN, MI, IL, WI, IA, and MN) and the northeastern states (NY , MA, CN, VT, NH, and ME) and the southeastern provinces of Canada (Ontario, New Brunswick, Quebec, and Nova Scotia). Peak ozone concentrations in the SOS region are generally lower than those in the mid-Atlantic, northeastern states, and some of the eastern provinces of Canada. But minimum concentrations of ozone are generally higher. Furthermore, ozone accumulation in the SOS region is decoupled from ozone accumulation in the mid-Atlantic and northeastern states. The differences between these three regions are due in part to the greater frequency of weather-front passages in the mid-Atlantic and northeastern states and the greater frequency of air-stagnation events in the southeastern states (Vukovich, 1992, 1994).

C2. In contrast to the northeastern states, ozone episodes in the SOS region are characterized by regionally dispersed, but spatially incoherent areas of high ozone concentration punctuated on the mesoscale by 'hot spots' of high ozone concentrations (Chameides and Cowling, 1995).

C3. Ozone concentrations throughout the SOS region are high enough (in excess of 60 ppbv) to inhibit photosynthesis of crops, forest and shade trees, and other plants during some portion of the growing season in essentially every year (Heck and Cowling, 1997).

C4. Substantial year-to-year and month-to-month variability in daily maximum ozone concentrations was observed within the SOS region, and was attributed to climate fluctuations and variation in emissions, respectively (Vukovich, 1998).

C5. The spatial variability of ozone concentrations in the SOS region suggests that multiple monitoring sites in urban, suburban, and rural sites may be necessary to detect maximum ozone concentrations in a reliable way (Imhoff and Valente, 1995).

C6. Ozone concentrations in the SOS region are positively correlated with temperature and negatively correlated with amount of precipitation (Vukovich, 1994).

C7. On a climatological scale, interannual variations in either temperature or cloud cover explained about 80 percent of the variability in daily maximum ozone concentrations during the time period 1981-1990 (Vukovich, 1998).

C8. Multiple-regression models of ozone concentration with air temperature, wind speed, relative humidity, and ozone concentration during the previous 24 hours can provide a useful method for decreasing the effect of meteorological variability on the year-to-year ozone concentration trends (Vukovich, 1994).

C9. In examining short-term (1-5 days) ozone episodes, the most persistent relationship between ground-level ozone concentrations and weather parameters was between ozone concentrations and wind speeds, with stagnation periods leading to the highest daily maximum ozone concentrations. When a 15-year-long time series of ozone concentrations was compared with the same 15-year-long time series of meteorological patterns, however, the most persistent relationship was between ozone and cloud cover. When days with ozone concentrations equal to or greater than 100 ppb were extracted from the 15-year time series and examined separately, only wind speed and cloud cover were important. Neither temperature nor dew point was important on these high ozone days (Vukovich, 1998).

C10. Regional NOx and/or VOC emission control strategies may decrease the frequency of ozone exceedance events, but episodic NOx and/or VOC control strategies probably will be necessary to eliminate exceedance events completely (Vukovich, 1997).

These scientific findings (C1-C10 above) suggest that while the ozone problems in the ten SOS states and the mid-Atlantic and northeastern states and southeastern provinces of Canada have some common features, e.g., correlation of peak ozone with temperature and stagnation conditions, there also are significant differences that suggest application of different control strategies in the SOS region than in some other parts of the eastern US and southeastern Canada. Thus, concern in the northeastern and mid-Atlantic states and southeastern Canada logically should focus more often on short-term ozone episodes that generally are confined within urban areas and their effects on human health. In contrast, concern in the SOS region logically should focus more often on both short-term urban ozone episodes and also on the high and pervasive regional ozone concentrations and the effects on vegetation when ozone concentrations exceed 60 ppb). In terms of control strategies, emission controls in the mid-Atlantic and especially the northeastern states are justifiably limited mostly to NOx and VOC sources within the non-attainment area, whereas the situation in the SOS region suggests application of controls on both regional and urban scales.

C11. From 1982 to 2001, US national average ambient one-hour ozone concentrations decreased by about 18 percent and the corresponding 8-hour average ozone concentrations decreased by about 11 percent (USEPA, 2002). Also, one-hour exposures decreased for the nation as a whole from 1982 to 2001 on average by about 18 percent. The largest 20-year-trend decreases in ozone concentrations were observed in the northeastern states and the far western states (24-32 percent) and the smallest decreases were observed in the southeastern states and mid-Atlantic states (7-10 percent) (USEPA, 2002).

This scientific finding (C11 above) underscores the seriousness of the ozone problem in the SOS region and the need for a comprehensive research program addressed specifically to ozone problems in this region.

C12. From 1982 to 2001, estimated annual emissions of VOC in the United States as a whole decreased by about 16 percent, but estimated annual emissions of NOx increased in the United States as a whole by about 9 percent (USEPA, 2002).

Given the dual role of NOx in both the formation and in the destruction of ozone in the atmosphere, it is not clear whether it was the decrease in VOC emissions or the increase in NOx emissions that caused the downward trend in ozone concentrations cited in C13 above.

C13. No significant improvement in total exposure to ozone was observed in either rural areas or urban areas of the SOS region between 1980 and 1992 (Vukovich, 1994; Meagher and Parkhurst, 1996).

This 12-year period was one of substantial growth in human population and vehicle use in the SOS region, but was also a period of substantial investment in VOC controls and cleaner (lower emission) vehicles. It is discouraging that no significant decrease in average ozone concentrations occurred despite these investments, but it is also encouraging that there were no ozone increases caused by the increased growth. Evidently, the increased growth in human population and vehicle use and the increased investment in ozone controls during that period have substantially offset each.

C14. Air concentrations of reactive nitrogen (NOy) (i.e., unreacted and reacted NOx; or, to be more specific, mainly NO, NO2, nitric acid, PAN, organic and inorganic nitrates, and occasionally nitrous acid) generally are relatively low in rural parts of the SOS region - usually less than 4 ppbv (Kleinman et al., 1994).

Because NOy concentrations generally were below 4 ppbv in rural areas of the SOS region, and also because both ozone and NOy concentrations decrease with decreasing NOx, concentrations, this scientific finding (C15 above) suggests that ozone production in rural parts of the SOS region generally are NOx-limited. Thus, the optimum strategy for decreasing regional ozone accumulation in the SOS region is control of those NOx sources that impact the SOS region's rural areas.

C15. The adverse effects of ozone on agricultural crops, forest and shade trees, and other natural vegetation when ozone concentrations exceed 60 ppbv indicate that a secondary (welfare-based) National Ambient Air Quality Standard (NAAQS) for ozone - a standard that would be different in form from the present one-hour or eight-hour primary (human health-based) NAAQS - would provide an increased margin of safety for agricultural crops, forest and shade trees, and natural vegetation against the injurious and economic-damaging effects of ozone on ecosystems. The often-proposed SUM06 secondary standard for ozone was the consensus choice recommended by ecologists who participated in an SOS workshop on a possible secondary standard for ozone (Heck and Cowling, 1997; Chameides et al., 1997).

Adoption of a secondary standard for ozone (such as the often proposed SUM06 standard) with a more moderate ozone concentration (e.g., 60 ppb) but a longer 3-month or growing-season long average time will cause large portions of the rural parts of the SOS region and other states to be designated non-attainment areas. These possible changes also will tend to shift the major focus of concern about ozone effects from public health alone, to human and ecological health. It also will tend to shift the major concern about precursor emission controls from urban areas alone, to include also both suburban and rural areas, as is already the case in the European Union.

C16. The ozone concentration data in EPA's Aerometric Information Retrieval System (AIRS) are largely inadequate for identifying rural areas that would be in non- attainment if a welfare-based secondary standard such as SUM06 were adopted by the USEPA or any of the states in the US or eastern provinces of Canada. More appropriate data, covering rural areas, are those taken in SOS' Spatial Oxidants Network (SON) and EPA's Clean Air Status and Trends Network (CASTNet) (Chameides et al., 1997).