Ozone exposures in rural Maryland's Lower Eastern Shore.

Introduction

Ground level ozone, a powerful 'oxidant, is a main component of smog and the most toxic of the six "criteria air pollutants" regulated by the U.S. Environmental Protection Agency (U.S. EPA) (1). More people in the United States are exposed to ozone than to any other ambient air contaminant (2). During 1991-1993, approximately 55 percent of the U.S. population resided in counties and cities in which ozone concentrations exceeded levels that produce adverse human health effects for eight or more hours at least four times a season (2). In 1994, approximately 50 million people resided in U.S. counties in which ozone levels were above the current National Ambient Air Quality Standard (NAAQS) of 0.120 ppm (3). On May 31, 1996, an Advance Notice of Proposed Rulemaking by U.S. EPA proposed revising the existing NAAQS and focused on several lower standards (4). On July 18, 1997, the Final Rule for ozone was promulgated; it specified an eight-hour-average limit of 0.08 ppm (5).

Most studies of monitoring data from rural areas (except rural areas immediately downwind of large urban areas) have focused on the frequency distribution and seasonal and diurnal variations of ozone rather than on rural population exposure (6). Because of the lack of sources that would generate large amounts of ozone precursors, most rural exposures are believed to be lower than most urban areas and to affect only a comparatively small population. Under any of the proposed revisions, including the new standard, the number of counties that exceed the regulatory air quality standard could have increased dramatically to include many rural areas that met the earlier regulatory standard. The lowered NAAQS could, therefore, increase the population of rural residents who would be considered at risk.

The lack of consistent air quality monitoring makes exposure of rural residents difficult to assess and potentially puts members of this population at a risk similar to that incurred by their urban counterparts. Moreover, assessments of air quality risk in rural areas often do not consider that these areas attract large numbers of tourists and summer weekenders who reside in the urban areas. These vacationers may spend the better part of their time outdoors in unmonitored areas where exposure risk for ozone and other air pollutants is unknown.

Monitoring in the study area, a rural area on the Eastern Shore of Maryland, was conducted from 1985 to 1988 by the state of Maryland to assess ozone levels [ILLUSTRATION FOR FIGURE 1 OMITTED]. The state discontinued monitoring in 1988 and moved the instrument to an urban site because the data were interpreted as indicating low risk. Since that time, studies on the physiological effects of ozone have led researchers to believe that there is no threshold for adverse health effects related to ozone (7). Exposure to low concentrations over a period of time can result in respiratory illness or tissue damage (79). The many other studies that demonstrate the adverse health effects of low-level exposure are summarized by U.S. EPA's Air Quality Criteria Document for Ozone and Related Photochemical Oxidants (10). In light of these findings, this study re-examined the data from the four years of monitoring in the study area with the goal of estimating current exposure for the resident and tourist population in this rural area of Maryland. In addition, the effect of a new ozone NAAQS on Lower Eastern Shore regulatory compliance is assessed, and the broader impact on the region's health, economy, and lifestyle is discussed.

Ozone Sources

Ozone generation is favored in densely populated areas, such as major urban centers, that have high volumes of traffic and heavy industries that generate smog. Precursors from these sources are converted to ozone after hours of exposure to high temperatures and intense sunlight under conditions of low wind speeds and no precipitation (6). These conditions are most likely to occur during the late spring and summer. Also, there is evidence that ozone in rural areas can be generated biogenically from volatile hydrocarbons released from vegetation (11). This effect is most intensive under the same weather conditions that generate ozone in urban areas. Elevated ozone levels in nonurban areas are thought, however, to result mainly from atmospheric transport of ozone and ozone precursors from upwind sources such as large urban centers (6). The area considered in this study is ringed on three sides by major metropolitan areas that could serve as ozone sources [ILLUSTRATION FOR FIGURE 1 OMITTED].

Health Effects

In June 1996, a study of 13 U.S. cities by researchers from the Harvard School of Public Health found that ozone may be responsible for as many as 50,000 hospital emergency room visits annually in this country (12). During the 1994 summer season, 8,000 people were admitted to Washington, D.C., hospitals with respiratory illnesses. The Harvard researchers estimated that 7.6 percent of the admitted cases, or 600 admissions, were for illnesses resulting from exposures to elevated ozone levels. In Baltimore, the investigators estimated that 7.9 percent of admissions were associated with ozone exposure. Washington and Baltimore are the major cities nearest to the study area, so unhealthful concentrations of ozone are present and available for transport throughout the region. No data are readily available on the number of people affected by ozone in rural Maryland.

The health effects of ozone are mainly respiratory effects that occur with exposure to the irritant gas (7). According to the Harvard study, pneumonia was the most common cause of hospital admissions related to elevated ozone levels, while asthma was the second most common (12). Other respiratory problems requiring hospital admission included bronchitis and influenza. Once inhaled, ozone may oxidize and inflame lung tissue, cause chest pain, increase susceptibility to infections, and precipitate episodes of asthma and allergies. Subpopulations that are particularly sensitive to ozone pollution (known as "responders" or "sentinel populations") include children, the elderly, and people with pre-existing respiratory conditions. In asthmatics and other individuals with chronic obstructive pulmonary disease (e.g., emphysema), respiratory effects resulting from ozone may be severe.

Unlike environmental damage from this pollutant, which appears to be related to peak concentrations, respiratory effects may occur with exposure to low levels over several hours (13,14). Pulmonary effects appear to be cumulative, resulting in premature aging of the lung tissue with continuing exposures. Permanent structural damage may occur with months to years of ozone exposure.

TABLE 1

Descriptive Statistics for Daily Ozone Maxima and Eight-Hour Daytime
Averages (1985-1988 Ozone Seasons)

            n       Mean (ppm)       SD         CV (%)      SE Mean

Maxima

1985       214      0.066           0.0204       30.8       0.00139
1986       202      0.060           0.0233       38.9       0.00164
1987       198      0.046           0.0137       29.9       0.00098
1988       169      0.046           0.0175       38.0       0.00134

TOTAL      783      0.055           0.0210       38.1       0.00075

Eight-Hour Means

1985       214      0.051           0.0171       33.5       0.00117
1986       202      0.045           0.0195       43.4       0.00137
1987       198      0.035           0.0110       31.3       0.00078
1988       169      0.035           0.0149       42.7       0.00122

TOTAL      783      0.042           0.0174       41.3       0.00062

[TABULAR DATA FOR TABLE 2 OMITTED]

Regulatory Standards

The Clean Air Act of 1970 requires that U.S. EPA set a health standard that protects public health with an "adequate margin of safety" (15). U.S. EPA is also required to reexamine health standards every five years. Many of the health-related studies of ozone conducted since 1986 indicate the potential for adverse health effects at ozone levels below the current standard and over exposure periods longer than one hour. These new findings are the reason that U.S. EPA has considered a lower standard (16). Because of the absence of a threshold for health effects, U.S. EPA asserts that there is no minimum level (or "zero-risk" standard) that would necessarily eliminate all risk. Under the new rule, rural areas such as the study area may find that they exceed the standard much more often than they do at present.

The Clean Air Science Advisory Committee (CASAC) of the U.S. EPA Science Advisory Board recommended a range of 0.07 to 0.09 ppm for the new standard. U.S. EPA's Final Rule for ozone falls within that recommended range and replaces the one-hour primary standard with a new eight-hour standard of 0.08 ppm to protect against the cumulative health effects of longer periods of exposure to ozone concentrations below the current standard. The new eight-hour standard looks at the maximum eight-hour concentration for each day, takes the fourth-highest eight-hour maximum that occurs each year for three years, then averages those three values. If the result is 0.08 ppm or less, the area is in compliance with the standard. In this way, the standard can target ozone concentrations associated with the cumulative health effects of low-level exposure.

Study Location

The Lower Eastern Shore of Maryland is a primarily rural region composed of four counties located on the Delmarva Peninsula just to the north of the Virginia portion of the peninsula [ILLUSTRATION FOR FIGURE 1 OMITTED]. The region contains 4,854 square kilometers (1,896 square miles). Its total population as of 1990 was 163,043 with a population density of 33.6 people per square kilometer, which is only slightly greater than the average density for the whole United States of 30 people per square kilometer. By census definition, the rural population was 108,643, indicating that 67 percent of the population is considered rural. The urban population was 54,400, of which 37,255 (68 percent of the urban population) are accounted for by the three largest cities in the region (Salisbury, Cambridge, and Ocean City).

The monitoring site was located in a rural part of the region on a state agricultural experimental farm [ILLUSTRATION FOR FIGURE 1 OMITTED]. Because the ozone monitor was located in a rural environment far from major urban centers, its measurements reflect the region's rural ozone concentrations. Salisbury, with a 1990 population of 20,592, is the largest city in the area and is located approximately 10 km southeast of the monitoring site. The second most populous urban entity, Cambridge (11,514) is located approximately 40 kilometers northwest of the site. The third-largest city in the area is Ocean City. Its permanent population of 5,146 can swell to more than 250,000 on weekends during the tourist season, which also corresponds to the ozone season. Consequently, if ozone exposure is a problem in the study area, many more people may be affected than indicated by census counts. The other population concentrations are small (generally fewer than 500) and are found throughout the area. The major land uses in the study area are forest and agriculture (11).

Methods

Ground level ozone data for the study area were obtained from the Federal Aerometric Information Retrieval System (AIRS) and consisted of hourly ozone concentration averages for each day from March through October for 1985 and 1986 and from April through September for 1987 and 1988. The data for the study period were collected by the SLAMS in rural Wicomico County, Maryland. The official ozone season in Maryland is April 1 through October 31 (17). Only days on which the SLAMS monitor recorded complete hourly data sets (24 hours) were considered - or a total of 1,006 days. From April 1985 through March 1986, data were obtained with instrumental chemiluminescence. From April 1986 through November 1988, data were obtained with an instrumental ultraviolet method. The probe height was 5 meters above ground level. The minimum detectable ozone concentration for all data was 0.005 ppm. Both methods are standard and produce reliable estimates of ozone concentrations, but instrumental ultraviolet methods are much easier and have largely replaced instrumental chemiluminescence.

Descriptive statistics were computed for the data set, including the determination of the data's statistical distribution. Averages were calculated for each season, as well as for the whole data set, to determine background ozone exposures. Standard deviations were calculated for each mean to describe the variability in the data. The statistical distribution was used to analyze the overall form of the data set. Also tabulated were the numbers of hourly averages that exceeded 0.07, 0.08, and 0.09 ppm each day. This approach allowed comparisons of actual rural ozone concentrations from the data set with each of several values permissible under the various standards proposed by CASAC. Additionally, the data set was examined to determine whether historic values would have caused Wicomico County to fail to meet the new eight-hour standard (which is met when the three-year average of the annual fourth-highest maximum eight-hour concentration is 0.08 ppm or less).

Results

Daily Maxima

The mean and standard deviation of the daily ozone concentration maxima (the highest one-hour mean) for the Maryland ozone season (April 1-October 31) for the four years were 0.055 ppm and 0.021 ppm, respectively (n = 783) (Table 1). Maximum values ranged from a low of 0.015 ppm to 0.152 ppm. Like daytime averages, ozone maxima varied with season. Mean daily ozone maxima were well above the study period average during the 1985 and 1986 seasons (0.066 and 0.06 ppm, respectively). In the 1987 and 1988 seasons they were much lower (0.046 ppm for both years).

Eight-Hour Mean

The average daytime (8 a.m. to 8 p.m.) ozone concentration during the ozone season (April through October) was 0.042 ppm, with a standard deviation of 0.017 ppm (n = 783). The 1985 and 1986 averages were 0.051 and 0.045 ppm, respectively, while in both 1987 and 1988 the average was a much lower 0.035 ppm, indicating significantly better air quality during the latter two seasons (F = 49.94; p [less than] .001) (Table 1).

The standard deviations for both the eight-hour and daily-maximum means reflect the variability in the data and suggest that these mean values are probably less important than seasonal and even daily levels. A large variance in the data is not surprising given the data range. Consequently, these averages are not a reliable measure of typical exposure in the study area. Ozone can be a much greater hazard than indicated by average values and, at times, can warrant much stronger actions than those recommended by the three-year average-concentration standard. Furthermore, the brevity of the historical record recommends caution before it is assumed that either the first two years or the second two years are representative. None of the averages reflect the many excursions above health-significant levels shown in Table 2 and Table 3. In addition, the 0.120 ppm hourly ozone standard in effect at the time was exceeded eight times during the study period. Thus, the Air Quality Index was 100+ (greater than 100 percent of the 0.120 ppm standard) - indicating "Code Red" conditions and the presence of unhealthy air - eight times (18). All eight occurrences were recorded in 1985 and 1986.

Number of Times per Year that Ozone Concentrations Exceeded Various
U.S. EPA-Suggested Hourly Standards (1985-1988)

           Proposed Standard        Previous Standard
                (ppm)                     (ppm)

Year     0.07     0.08     0.09           0.120

1985     572      227       87              1
1986     445      235      104              2
1987      22        0        0              0
1988      93       17        2              0

[TABULAR DATA FOR TABLE 4 OMITTED]

Frequency Distribution

The data were examined to determine the statistical distribution of ozone concentrations over the four years. These data are not normally distributed (R = 0.9895, p [less than] .0100). Also, unlike the typical distribution of environmental data sets, our data do not follow a log-normal distribution (R = 0.9929, p [less than] .0100). With common statistical transformations, the best fit was found to be a square root transformation (i.e., the square root values of the daily ozone maxima are normally distributed) (R = 0.9990, p [less than] .0418). The squared values of the square root mean and standard deviation of the maxima were 0.059 ppm and 0.002 ppm, respectively. A square root distribution may suggest a combination of data sets with large differences in both means and standard deviations because they result from different mechanisms of ozone occurrence. When proportional to the means, the variances can be stabilized by a square root transformation (19).

Projected Occurrences of Ozone Levels that Would Have Exceeded Hourly Standards

To assess the impact of the various NAAQS ozone standards that were considered, the number of times ozone levels from the historical data exceeded the standard was determined for 0.07, 0.08, and 0.09 ppm one-hour standards (Table 3). In contrast to the eight occurrences in 1985 and 1986, the projected number of times ozone levels would have exceeded the standard ranged from 572 under a 0.07 ppm standard to 87 for a 0.09 ppm standard. During 1987 and 1988, which were low-ozone years, the number of occurrences ranged from zero for the 0.08 and 0.09 ppm standards to 93 for the 0.07 ppm standard.

Projected Numbers of Days on Which the New Eight-Hour Standard Would Have Been Exceeded

Under the new 0.08 ppm. eight-hour standard, projections for the data set are comparable to those found for the 0.09 ppm one-hour suggested standard; the number of times ozone levels exceeded the standard was much higher for 1985 and 1986 than for 1987 and 1988 (Table 2). The regulatory impact of the two standards on the area would have been virtually identical during the study period.

Projected Compliance

The new 0.08 ppm, eight-hour standard is met when the fourth-highest annual maximum eight-hour concentration, averaged over three years, is equal to or less than 0.08 ppm. Table 4 shows that the study site would have been in compliance from 1985 to 1987. The critical value for those three years was 0.078 ppm, or, rounded to the two significant digits, 0.08 ppm.

Urban Maryland Versus Rural Eastern Shore

Ozone seasons for the study years were compared with those for the urban centers across the Chesapeake Bay. For sites .in urban Maryland, the year 1988 had the highest ozone levels since 1983; 1985 and 1986 had only moderate levels. However, the study area showed extremely low levels in 1988 and moderate to high levels in 1985 and 1986.

Discussion

Stakeholders responsible for air quality in the United States are aware of elevated ozone in rural regions, but there is little hard evidence that ozone exposure is a problem for people residing in the countryside (20). Routine monitoring in nonurban areas has not been a priority, especially in times of fiscal downsizing. The analysis of Lower Eastern Shore data, however, indicates that at least this rural area is subject to levels of ozone that significantly affect air quality and pose potential health risks. The contrast between the numbers of days and hours of unhealthy air quality and the achievement of regulatory compliance is especially striking (Tables 1, 2, and 4). Results suggest that rural areas can be subject to levels of ozone exposure that justify long-term monitoring to determine air quality conditions and measure exposure levels, regardless of compliance status.

Suburban and even some rural areas contiguous to a nearby source are clearly vulnerable to ozone transport and may themselves generate considerable ozone. The study area and many other rural areas that are distant from large urban sources generate little ozone from local urban activities anti seem far removed from air quality alerts and increased hospital admissions. The study area, however, is bounded on the north, south, and west by very large urban centers. Meteorological transport may carry ozone from multiple urban sources at different distances and in different directions. Depending on wind direction and strength, as well as local weather conditions at each ozone source, the amount of ozone and ozone precursors transported to rural areas may vary greatly.

The differences between the ozone concentrations of the Baltimore and Washington urban centers and those of the rural Maryland study area seem to indicate that rural ozone levels can vary independently from urban sources. The differences also suggest that the mechanisms that control ozone occurrence vary with geography, even over the relatively short distance between the Baltimore-Washington area and the Lower Eastern Shore ([+ or -]200 kilometers). The use of urban monitoring sites could therefore yield inaccurate predictions of ozone concentrations in rural areas. Yet the only ozone information available to residents in rural Maryland is based on air quality monitors in the Baltimore-Washington urban corridor.

Apparently, regional weather patterns in 1987 and 1988 favored ozone generation and transport along the Baltimore-Washington corridor but did not induce ozone in the study area. Another consideration is the moderate level of biogenically produced volatile organic compounds (e.g., terpenes from local pine forests), which could constitute significant numbers of ozone precursors and thus an additional source of ground level ozone in the study area (11). The mechanisms that control biogenic ozone production are not well known, and the local emission rates are therefore estimated; however, those rates would be expected to vary with local environmental conditions. For all these reasons, the sources and mechanisms of ozone occurrence and, consequently, ozone concentrations in the study area could differ significantly from those of a single urban source. The difference argues for local monitoring sites in rural areas such as the Maryland study area, despite the fact that ozone concentrations in general are lower there than in urban areas.

While ozone monitoring in rural areas could result in greater protection against the health effects of ozone pollution, increases in the number of times standards are exceeded and in the number of air quality alerts could have unintended costs. During periods of poor air quality, it is strongly recommended that outdoor activity be restricted for children, the elderly, and people with respiratory problems; even healthy individuals need to restrict their outdoor activities (7). On the Lower Eastern Shore, tourism is an important economic activity. Many retirees from the surrounding urban centers have begun to settle in the area because of its rural character. Therefore, actions that restrict outdoor activity could seriously reduce the attractiveness of the area for tourism and retirement, which would affect the local economy and its future growth.

Furthermore, an increase in the number of times standards are exceeded could trigger regulatory mandates, which often require increased compliance spending such as enhanced inspection-maintenance programs for automobiles, installation of more stringent emissions controls, and expensive retrofitting of equipment. Unfortunately, if much of the ozone is transported from urban areas or is biogenically derived, these remedial activities will not significantly reduce ozone levels in the rural area - but they may still be mandated.

Conclusions

The pattern of exposure to ozone during the study period varied from year to year (and from month to month within each year) to the extent that investigators were unable to determine the "typical" annual ozone exposure from the available data. Without additional monitoring, it is not possible to know whether the residents in the study area are exposed to excessive ozone levels with some regularity or whether the years of relatively high ozone levels were anomalous. We cannot, however, ignore past and potential future exposure. A review of these data clearly indicates that an extrapolation from monitoring performed across the Chesapeake Bay in the urban areas of Washington, Annapolis, and Baltimore could underestimate ozone levels in the study area.

Local monitoring would have several benefits, including a more accurate assessment of local ozone conditions that could (and should) be provided to rural residents for effective public health education, the determination of environmental conditions under which local ozone levels are high, and a better possibility that the exact mechanism that controls local ozone levels can be identified. The determination of that mechanism can lead to the best possible remediation effort: the reduction of ozone concentrations at its sources.

Acknowledgements

As the Air Quality Research Group, the authors have established a monitoring site at Salisbury State University with the encouragement and active support of the Air and Radiation Management Administration of the Maryland Department of Environment, the Maryland Department of Natural Resources, and Delmarva Power and Light. The authors gratefully acknowledge the support of these entities. Data from the site will be used to determine current exposure to ozone in the study area.

The authors also gratefully acknowledge the support and encouragement of the Richard A. Henson School of Science and Technology and Office of Grants and Sponsored Research of Salisbury State University, which made the establishment of this monitoring site possible.

Finally, the authors thank Kathy D. Pusey, Salisbury State University, Public Relations, for the use of her photographs in the manuscript.

REFERENCES

1. Godish, T. (1997), Air Quality, 3rd ed., Boca Raton: Lewis Publishers.

2. Centers for Disease Control and Prevention (1995), "Children at Risk from Ozone Air Pollution - United States, 1991-1993," Morbidity and Mortality Weekly Report. 44(16):309-312.

3. Fact Sheet: Health and Environmental Effects of Ground-Level Ozone (1990). Washington, D.C.: U.S. Environmental Protection Agency, Office of Air and Radiation.

4. "Advance Notice of Proposed Rulemaking, (May 31, 1996), 61 Federal Register.

5. "National Ambient Air Quality Standards for Ozone: Final Rule," (July 18, 1997), Federal Register; 62:138; 40 CFR [section]50.

6. National Research Council (1991), Rethinking the Ozone Problem in Urban and Regional Air Pollution, Washington, D.C.: National Academy Press.

7. Costa, D.L., and M.O. Amdur (1996), "Air Pollution," In Casserette and Doule's Toxicology - The Basic Science of Poisons, 5th ed., C.D. Klaasen, ed., New York: McGraw-Hill.

8. Bates, D.V. (1995), "Ozone: A Review of Recent Experimental, Clinical and Epidemiological Evidence. with Notes on Causation: Parts 1 and 2," Canadian Respiratory Journal, 2(Spring):1.

9. Lippman, M. (1993), "Health Effects of Tropospheric Ozone: Review of Recent Research Findings and Their Implications to Ambient Air Quality Standards." Journal of Experimental Analysis and Environmental Epidemiology, 3(1): 103-129.

10. Office of Health and Environmental Assessment, Environmental Criteria and Assessment Office (1993), Air Quality Criteria for Ozone and Related Photochemical Oxidants, EPA/600/AP-32/004a-c, Research Triangle Park, N.C.: U.S. Environmental Protection Agency.

11. 1995 Ozone Atlas for the Mid-Atlantic Region (1997), Baltimore, Md.: Mid-Atlantic Regional Air Management Association.

12. Ozkaynak, H., K Gilbert, M. O'Neill, S. Ramstrom, J.D. Spengler, J. Xue, and H. Zhou (1996), Ambient Ozone Exposure and Emergency Hospital Admissions and Emergency Room Visits for Respiratory Problems in Thirteen U.S. Cities, Washington, D.C.: American Lung Association.

13. Adams, R.M., S.A. Hamilton, and B.A. McCarl (1985), "An Assessment of the Economic Effects of Ozone on U.S. Agriculture,"Journal of the Air Pollution Control Association, 35:938-943.

14. Adams, R.M.,J.D. Glyer, S.L. Johnson, and B.A. McCarl (1989), "A Re-Assessment of the Economic Effects of Ozone on U.S. Agriculture," Journal of the Air Pollution Control Association, 39:960-968.

15. Clean Air Act (1090), 42 USC [section]7408-7409.

16. "National Ambient Air Quality Standards for Ozone Proposed Decision." (December 13, 1996), Federal Register; 61:241; 40 CFR [section]50.

17. "Protection of the Environment" (1995), 40 CFR 58, Appendix D, p. 170

18. Air and Radiation Management Administration (1992), Maryland Air Quality Data Report 1992, Baltimore: Maryland Department of the Environment.

19. Kleinbaum, D.G., L.L. Kupper, and K.E. Muller (1988), Applied Regression Analysis and Other Multivariable Methods, 2nd ed., Boston: PWS-Kent Publishing Company.

20. Lindsay, R.W., and W.L. Chamcides (1988), "High-Ozone Events in Atlanta. Georgia, in 1983 and 1984," Environmental Science and Technology 22:426-431.

Corresponding Author: Elichia A. Venso, Director and Associate Professor, Environmental Health Science Program, Richard A. Henson School of Science and Technology, Salisbury State University, Salisbury, MD 21801.