Of the six major air pollutants for which National Ambient Air Quality Standards (NAAQS) have been designated under the Clean Air Act, the-most pervasive problem continues to be ozone,1 the most prevalent photochemical oxidant and an important component of "smog." The most critical aspect of this problem is the formation of ozone in and downwind of large urban areas where, under certain meteorological conditions, emissions of nitric oxide and nitrogen dioxide (known together as Nor.) and volatile organic compounds (VOCs) can result in ambient ozone concentrations up to three times the concentration considered protective of public health by the U.S. Environmental Protection Agency (EPA).
Major sources of VOCs in the atmosphere include motor vehicle exhaust, emissions from the use of solvents, and emissions from the chemical and petroleum industries. In addition, there is now a heightened appreciation of the importance of reactive VOCs emitted by vegetation. NOx comes mainly from the combustion of fossil fuels; major sources include motor vehicles and electricity generating stations.
The occurrence of ozone concentrations that exceed the NAAQS in various
1The scientific community now has strong reason to believe that, unlike stratospheric (i.e., high-altitude) ozone concentrations, which are declining, concentrations of tropospheric (i.e., near-ground) ozone are generally increasing over large regions of the United States.
regions of the United States indicates that many people may be exposed to concentrations of ozone that EPA has determined to be potentially harmful. EPA reported that in 1989, about 67 million people lived in areas where the second-highest ozone concentration, a principal measure of compliance, exceeded the NAAQS concentration. Despite considerable regulatory and pollution control efforts over the past 20 years, high ozone concentrations in urban, suburban, and rural areas of the United States continue to be a major environmental and health concern.
The nationwide extent of the problem, coupled with increased public attention resulting from the high concentrations of ozone over the eastern United States during the summer of 1988, adds to the urgency for developing effective control measures. As the ozone attainment strategy presented in the 1990 amendments to the Clean Air Act is put into effect, the success of efforts to control the precursors of ozone will be of vital concern to Congress, to governmental regulatory agencies, to industry, and to the public.
The Charge To The Committee
The Committee on Tropospheric Ozone Formation and Measurement was established in 1989 by the Board on Environmental Studies and Toxicology of the National Research Council (NRC) in collaboration with the NRC's Board on Atmospheric Sciences and Climate to evaluate scientific information and data bases relevant to precursors and tropospheric formation of ozone, and to recommend strategies and priorities for filling critical scientific and technical gaps in the information and data bases. The committee's members had expertise that included atmospheric chemistry, measurement, mathematical modeling, pollution trends monitoring, transport meteorology, exposure assessment, air-pollution engineering, and environmental policy. The committee was specifically charged to address
• Emissions of VOCs (anthropogenic and biogenic2) and NOx;
• Significant photochemical reactions that form ozone, including differences in various geographic regions;
• Effects of precursor emissions on daily patterns of ozone concentration;
• Ambient monitoring techniques;
2''Anthropogenic emissions'' refers to emissions resulting from the actions of human society. "Biogenic emissions" refers to natural emissions, mainly from trees and other vegetation.
• Input data and performance evaluations of air-quality models;
• Regional source-receptor relationships;
• Statistical approaches in tracking ozone abatement progress;
• Patterns of concentration, time, and interactions with other atmospheric pollutants.
The committee was not charged to evaluate and did not address the adequacy of the NAAQS for protecting human health and welfare, nor the technologic, economic, or sociologic implications of current or potential ozone precursor control strategies.
The committee's work was sponsored principally by the U.S. Environmental Protection Agency and Department of Energy. Additional funding was provided by the American Petroleum Institute and the Motor Vehicle Manufacturers Association. EPA is expected to provide the committee's report to Congress as partial fulfillment of Section 185B of the Clean Air Act amendments of 1990, which requires EPA to conduct a study in conjunction with the National Academy of Sciences on the role of ozone precursors in tropospheric ozone formation and control. The study required in Section 185B is more extensive than the committee's charge, and EPA is separately addressing the portions of the study required in Section 185B that are beyond the scope of this NRC study.
The Committee's Approach to its Charge
In this report, the committee examines trends in tropospheric ozone concentrations in the United States; reviews current approaches to control ozone precursors; and assesses current understanding of the chemical, physical, and meteorological influences on tropospheric ozone. Based on this understanding, the committee provides a critique of the scientific basis for current regulatory strategies and offers recommendations for improving the scientific basis for future regulatory strategies. It also recommends an integrated research program to further clarify the factors that affect tropospheric ozone formation within the context of changing regional and global environmental conditions. The committee presents data and arguments to address these issues and to help resolve the continuing national debate over devising an effective program to achieve the NAAQS for ozone.
The major findings and recommendations of the committee are discussed in the remainder of this summary.
Ozone In The United States
FINDING: Despite the major regulatory and pollution-control programs of the past 20 years, efforts to attain the National Ambient Air Quality Standard for ozone largely have failed.
DISCUSSION: Since passage of the 1970 Clean Air Act amendments, extensive efforts to control ozone have failed three times to meet legislated deadlines for complying with the ozone NAAQS. Congress set 1975 as the first deadline, but 2 years after this deadline, many areas were still in violation of the NAAQS. The 1977 amendments to the Clean Air Act extended the deadline for compliance until 1982 and allowed certain areas that could not meet the 1982 deadline until 1987. For 1987, however, more than 60 areas still exceeded the NAAQS; the following year, the number of areas exceeding the NAAQS jumped to 101. In 1990, 98 areas were in violation of the NAAQS.
EPA has reported a trend toward lower nationwide average ozone concentrations from 1980 through 1989, with anomalously high concentrations in 1983 and 1988. Ozone concentrations were much lower in 1989 than in 1988, possibly the lowest of the decade. However, since the trend analysis covers only a 10-year period, the high concentrations in 1983 and 1988 cannot be assumed to be true anomalies, nor can the lower concentrations in 1989 be assumed to be evidence of progress. It is likely that meteorological fluctuations are largely responsible for the highs in 1983 and 1988 and the low in 1989. Meteorological variability and its effect on ozone make it difficult to determine from year to year whether changes in ozone concentrations result from fluctuations in the weather or from reductions in the emissions of precursors of ozone. However, it is clear that progress toward nationwide attainment of the ozone NAAQS has been extremely slow at best, in spite of the substantial regulatory programs and control efforts of the past 20 years.
FINDING: The principal measure currently used to assess ozone trends (i.e., the second-highest daily maximum 1-hour concentration in a given year) is highly sensitive to meteorological fluctuations and is not a reliable measure of progress in reducing ozone over several years for a given area.
RECOMMENDATION: More statistically robust methods should be developed to assist in tracking progress in reducing ozone. Such methods should
account for the effects of meteorological fluctuations and other relevant factors.
DISCUSSION: Year-to-year meteorological fluctuations might mask downward ozone trends in cases where precursor emission controls are having the desired effect. Such fluctuations might also mask trends of increasing ozone or evidence that progress in reducing ozone has been slower than expected. Alternative statistical measures, discussed in Chapter 2, should be developed. These measures should not be mere statistical entities; they should bear some relation to the range of ozone concentrations considered harmful to human health and welfare. Support should be given to the development of methods to normalize ozone trends for meteorological variation. Several techniques that have shown promise for individual cities and regions could be useful on a national scale.
State Implementation Planning
FINDING: The State Implementation Plan (SIP) process, outlined in the Clean Air Act for developing and implementing ozone reduction strategies, is fundamentally sound in principle but is seriously flawed in practice because of the lack of adequate verification programs.
RECOMMENDATION: Reliable methods for monitoring progress in reducing emissions of VOCs and NOx must be established to verify directly regulatory compliance and the effectiveness associated with mandated emission controls.
DISCUSSION: The Clean Air Act places the responsibility for attaining the NAAQS on a federal-state partnership. EPA develops uniform NAAQS, and the states formulate and implement emission reduction strategies to bring each area exceeding the NAAQS into compliance.
The SIP process for air-quality management is based on the premise that emission reductions can be inferred directly from observed improvements in air quality. For some pollutants, such as sulfur dioxide (SO2) and carbon monoxide (CO), it has generally been possible to infer progress by tracking emission reductions through measurements of ambient concentrations of the pollutants. However, in the case of ozone, which is formed by highly complex and nonlinear reactions that involve VOC and NOx precursors, it is extremely questionable in most cases to conclude that emissions reductions have occurred solely on the basis of observed trends in ambient ozone concentrations.
EPA's approach to ozone control, originally developed in 1971, has relied largely upon unverified estimates of reductions in precursor emissions; EPA has not required systematic measurements of ambient precursor concentrations. Systematic measurements of NOx and VOCs are needed in addition to ozone measurements to determine the extent to which precursor emissions must be controlled and to verify the effectiveness of the control measures undertaken. Over the past two decades, the substantial reductions in ozone concentrations predicted to result from the VOC emission reductions in major urban centers have not occurred. Moreover, the limited data available on ambient concentrations of VOCs suggest that the actual VOC emission reductions have been smaller than estimated in the SIP process. The reasons for this failure are largely unknown.
Designing a strategy to control precursor emissions, and tracking that strategy's effectiveness in controlling a specific VOC or NOx, requires systematic VOC and NOx measurements in strategic locations. Until verification programs are incorporated into the SIP process, the use of unverified emission inventories3 in air-quality models will continue to involve considerable uncertainties in predicting changes in ozone concentrations resulting from emission controls.
Anthropogenic VOC Emissions
FINDING: Current emissions inventories significantly underestimate anthropogenic emissions of VOCs. As a result, past ozone control strategies may have been misdirected.
RECOMMENDATION: The methods and protocols used to develop inventories of ozone precursor emissions must be reviewed and revised. Independent tests, including monitoring of ambient VOCs, should be used by government agencies to assess whether emissions are indeed as they are represented by emissions inventories.
DISCUSSION: As discussed in Chapter 9, there is substantial evidence that the methods and protocols used to develop anthropogenic VOC inventories are flawed and do not account adequately for all types of sources, nor for
3An emission inventory is a data base containing estimated emissions from various sources in a specific area and period (for example, nationwide emissions of VOCs per year).
the magnitude of VOC emissions. Ambient measurements of VOC/NOx ratios, for instance, consistently yield ratios that are larger than would be expected from emissions inventories. In addition, measurements of VOC concentrations near roadways and in tunnels, as well as ambient measurements of specific VOCs in urban areas, indicate that VOC emissions from mobile sources have been underestimated in these inventories by a factor of two to four. The discrepancy in mobile-source emissions is probably a result of several factors in current emission models. It is likely that the fleets used in dynamometer testing to determine emission factors are not representative of on-road vehicles, that speed correction factors and estimates of evaporative emissions are inaccurate, and that the Federal Test Procedure does not adequately simulate actual driving behavior. Moreover, current Inspection and Maintenance (I/M) programs do not appear to be leading to the emission reductions anticipated.
The underestimation of VOC emission inventories presents two problems. One problem involves the limited effectiveness of older emissions-control technology. It now seems likely that mandated emissions controls in past years have not been as effective as EPA had estimated. Even if the reductions claimed are correct, their effect as a percentage of the total VOC inventory is clearly smaller than expected, because total anthropogenic VOC emissions were underestimated. Although VOC reductions of 25-50% have been claimed for some areas, it is likely that the actual reductions were in the 10-25% range.
A more profound problem resulting from underestimating VOC emissions is the implication for future emission controls. As discussed in Chapters 6 and 11, the relative effectiveness of VOC and NOx controls for reducing ozone in a particular area depends on the ambient VOC/NOx ratios in that area. At VOC/NOx ratios of about 10 or less, VOC control is generally more effective, and NOx control may actually be counterproductive. At VOC/NOx ratios of 20 or more, NOx control is generally more effective. The nation's ozone reduction strategy has been based largely on the premise that VOC/ NOx ratios in most polluted urban areas fall in the less-than-10 range, where VOC control is more effective than NOx control. Hence, a major upward correction in VOC emission inventories could indicate the need for a fundamental change in the strategy used to abate ozone in many geographic areas.
The fact that a fundamental change in the nation's ozone reduction strategy could be necessary after two decades of costly efforts indicates that there could be an even more fundamental flaw in the overall design of the strategy. The current design depends too much on the assumption that emissions inventories are accurate, and not enough on adequate checks and tests of these inventories. Independent tests of the inventories and alternative approaches,
such as tunnel studies and remote sensing, need to be added to the design so that errors in the inventories can be uncovered and corrections can be made. The checks should include thorough monitoring of precursor concentrations in nonattainment areas to verify the effectiveness of emission controls. In addition, air-quality models that use ambient data instead of emissions inventories to identify important precursor sources, to verify emissions algorithms, and to determine ozone precursor relationships could serve as useful checks of models that require the entry of accurate data from emissions inventories.
Bongenic VOC Emissions
FINDING: The combination of biogenic VOCs with anthropogenic NOx can have a significant effect on photochemical ozone formation in urban and rural regions of the United States.
RECOMMENDATION: In the future, emissions of biogenic VOCs must be more adequately assessed to provide a baseline from which the effectiveness of ozone control strategies can be estimated before such strategies are applied for a specific urban core or larger regions. Ambient measurements of concentrations and emission rates are needed to improve the accuracy of biogenic VOC inventories.
DISCUSSION: Measurements of the ambient concentrations of isoprene and other VOCs known to be emitted by vegetation, as well as estimates of total emissions from biogenic VOC sources, suggest that these compounds help foster episodes of high concentrations of ozone in urban cores and other areas affected by anthropogenic NOx. Moreover, in many rural areas, especially those in the eastern and southern United States, more biogenic VOCs than anthropogenic VOCs are oxidized in the troposphere. However, quantifying the role of biogenic VOCs in specific episodes of high ozone concentrations is difficult because of large uncertainties in inventories.
The biogenic VOC contribution is a background concentration that cannot be removed from the atmosphere by emission controls. If anthropogenic VOC emissions are reduced, this background concentration will become a larger and more significant fraction of total VOCs. The committee's analysis (see Chapter 8) suggests that in many urban cores and their environs, even if anthropogenic VOC emissions axe totally eliminated, a high background concentration of reactive biogenic VOCs will remain; for example, on hot summer days this can be the equivalent of a propylene concentration of 10-30 parts per billion (ppb) (see Chapter 8). In the presence of anthropogenic NOx and under favorable meteorological conditions, these background biogenic VOCs
can contribute to summertime ozone concentrations exceeding the NAAQS concentration of 120 ppb.
Ambient Air Quality Measurements
FINDING: Ambient air quality measurements now being performed are inadequate to elucidate the chemistry of atmospheric VOCs or to assess the contributions of different sources to individual concentrations of these compounds.
RECOMMENDATION: New measurement strategies that incorporate more accurate and precise measurements of the individual trace compounds involved in ozone chemistry should be developed to advance understanding of the formation of high concentrations of ozone in the United States and to verify estimates of VOC and NOx emissions.
DISCUSSION: There have been major advances in the ability to measure ambient VOCs, NOx, and other reactive nitrogen compounds (see Chapter 7). These more accurate and precise measurements should be used in coordinated programs to test the reliability of estimated VOC and NOx emissions from major anthropogenic and natural sources and to study the distribution of these sources to characterize ozone accumulation and destruction. Trends in atmospheric measurements of emissions can provide a valuable check of the reductions estimated by using emissions inventories.
In addition, despite significant progress in the past 2 decades, the state of knowledge of many of the fundamental processes that govern the formation and distribution of ozone in the atmosphere needs improvement. To increase understanding of the processes of ozone formation and removal, reliable measurements of ozone and its precursors are needed. These measurements can be obtained only by intensive field studies that use validated methods to provide data for evaluating the representation of physical and chemical processes in air-quality models designed to predict the effects of future emissions controls. Past field studies have offered limited spatial coverage; sites in a wide variety of areas must be studied. Future studies must consider winter as well as summer conditions. The role of aerosol particles and cloud chemistry in ozone formation must also be investigated.
FINDING: Although three-dimensional or grid-based ozone air-quality
models are currently the best available for representing the chemical and physical processes of ozone formation, the models contain important uncertainties about chemical mechanisms, wind-field modeling, and removal processes. Moreover, important uncertainties in input data, such as emissions inventory data, must be considered when using such models to project the effects of future emissions controls.
RECOMMENDATION: Air-quality models are essential in predicting the anticipated effects of proposed emissions controls on ambient ozone concentrations. Therefore, the effects of uncertainties on model predictions, such as uncertainties in the emissions inventory and in the chemistry incorporated in the models, must be elucidated as completely as possible. Predictions of the effects of future VOC and NOx controls should be accompanied by carefully designed studies of the sensitivity of model results to these uncertainties.
DISCUSSION: Most future SIPs will rely on trajectory and grid-based photochemical air-quality models to estimate the effects of strategies to control emissions of ozone precursors. Air-quality models are designed to represent the complex physics and chemistry of the atmosphere and require a number of important types of input data, such as initial and boundary conditions, meteorological fields, and emissions inventories (see Chapter 10). Models are evaluated for use in control strategy assessment by first establishing their ability to simulate one or more past episodes of high concentrations of ozone. Grid-based models generally have been able to simulate observed 1-hour average ozone concentrations with a gross error of 30% or less, but known biases in certain inputs, particularly in emissions inventories, have raised the concern that there are uncertainties as yet unknown in other areas such as meteorology. Thus, good predictions of ozone concentrations can result from offsetting errors in inputs. Predictions of the effect of future emission controls should be accompanied by estimates of the uncertainty about ozone concentrations that stems from uncertainties in input variables.
Uncertainties in air-quality models include those in the chemical mechanism, those in the treatment of physical processes, and those that result from the choice of numerical algorithms. One measure of the uncertainty arising from a chemical mechanism can be obtained by performing simulations with different chemical mechanisms. By necessity, simplified equations are used and constant values are assigned to represent physical processes in air photochemistry, such as atmospheric transport including dry deposition. The best approach to quantifying the effects of uncertainties in a model's representation of physical processes is to conduct simulations with a different model for the same set of input data.
Uncertainties in model input data include those involving meteorological factors (e.g., winds and mixing heights), initial and boundary conditions, base case emissions, and projected future emissions. The data that describe the wind field and mixing height contain uncertainties of magnitudes that are difficult to estimate. One way to quantify these uncertainties is to simulate different episodes for the same city or region. The overall influence of initial and boundary conditions on predictions of ozone concentrations can be assessed by performing simulations with initial conditions set to zero and boundary conditions set to zero. Similar calculations can be carried out to assess the uncertainties in ozone predictions that arise from uncertainties in base-year emissions inventories and projected future emissions.
Although considerable effort has gone into developing and applying air-quality models, the lack of ambient data that can be used to evaluate the models comprehensively has impeded progress in their development and use. To obtain such data will require intensive field programs, designed to obtain the data needed to evaluate models. The Southern California Air Quality Study (SCAQS), the San Joaquin Valley Air Quality Study (SJVAQS)/Atmospheric Utility Signatures, Predictions, and Experiments (AUSPEX), the Southern Oxidants Study, and the Lake Michigan Ozone Study are examples of such programs. More such programs should be developed, especially in the eastern and southern United States.
VOC Versus Nox Control
FINDING: State-of-the-art air-quality models and improved knowledge of the ambient concentrations of VOCs and NOx indicate that NOx control is necessary for effective reduction of ozone in many areas of the United States.
RECOMMENDATION: To substantially reduce ozone concentrations in many urban, suburban, and rural areas of the United States, the control of NOx emissions will probably be necessary in addition to, or instead of, the control of VOCs.
DISCUSSION: Application of grid-based air-quality models to various cities in the United States shows that the relative effectiveness of VOC and NOx controls in ozone abatement varies widely. NOx reductions can have either a beneficial or detrimental effect on ozone concentrations, depending on the locations and emission rates of VOC and NOx sources in a region. The effect of NOx reductions depends on the local VOC/NOx ratio and a variety of other factors. Modeling studies show that ozone should decrease
in response to NOx reductions in many urban areas. However, some modeling and field studies show that ozone concentrations can increase in the near field in response to NOx reductions, but decrease in the far field. Thus, NOx controls could reduce ozone under some conditions, but under different conditions might lead to smaller ozone decreases than if VOCs alone are reduced. NOx controls should be evaluated not only on the basis of reducing peak ozone concentrations, but also for their effects on other nitrogen-containing species (see Chapter 6). In any event, the ramifications of NOx control on ozone concentrations are complex and must be considered carefully in devising ozone abatement strategies.
A decrease in emissions of NOx should lower ozone concentrations in many parts of the United States. When rural measurements of VOCs and NOx are used in regional air-quality modeling, the resulting ozone concentrations can exceed 100 ppb-as is consistent with observed concentrations. In some areas, concentrations of the biogenic VOC isoprene by itself are high enough to generate this much ozone in the presence of ambient NOx. In rural areas, formation of ozone appears to be insensitive to changes in concentrations of anthropogenic VOCs because of the generally high VOC/NOx ratios, but this insensitivity depends strongly on how much NOx is present. Much of the NOx in rural areas is generated by mobile and stationary sources in the urban cores and their major connecting regions; such NOx sources appear to contribute to the pervasive high ozone concentrations found in the eastern United States. Simulations with the Regional Oxidant Model (ROM) have shown that ozone concentrations above 80 ppb can be generated in the synoptic-scale transport region, including the Ohio River Valley and the entire Northeast corridor, from the prevailing NOx and biogenic VOCs alone. Ozone concentrations were greater than 100 ppb downwind of the major urban areas.
Ozone is predicted to decrease in response to NOx reductions in most urban locations. Models show that ozone concentrations rise in some urban cores, such as New York City and Los Angeles, in response to NOx reductions but decrease in downwind areas, where maximum amounts of ozone are found. Choosing not to reduce NOx in those urban centers, while ameliorating a local problem, could exacerbate the ozone problem in downwind regions. Moreover, total population exposure to ozone (and other harmful pollutants) might not respond in the same manner as peak ozone to control techniques designed to reduce peak ozone; population exposure should be considered in the design of future control strategies.
Many simulations conducted to date have relied on emissions inventories that did not include biogenic emissions and are strongly suspected of significantly underestimating anthropogenic VOC emissions. The result is an overestimate of the effectiveness of VOC controls and an underestimate of the
efficacy of NOx controls. Faulty inventories have likely led to underprediction of ozone concentrations in central urban areas (see Chapters 10 and 11). An increase by a factor of two to three in mobile-source VOCs, as suggested by recent studies discussed in Chapter 9, leads to predicted ozone concentrations closer to those observed. If the anthropogenic VOC inventory is as badly underestimated as recent studies indicate, areas that were previously believed to be adversely affected by NOx controls might actually benefit from them.
Alternative Fuels for Motor Vehicles
FINDING: The use of alternative fuels has the potential to improve air quality, especially in urban areas. However, the extent of the improvement that might result is uncertain and will vary depending on the location and on the fuels used. Alternative fuel use, alone, will not solve ozone problems nationwide. Moreover, it will not necessarily alleviate the most critical problem associated with motor vehicle emissionsincreased emissions as in-use vehicles age.
RECOMMENDATION: Because there is uncertainty about the degree to which alternative fuels would reduce ozone, requiring the widespread use of any specific fuel would be premature. An exception may be electric vehicles, which can lead to substantial reductions in all ozone precursor emissions. Coordinated emissions measurement and modeling studies should be used to determine which fuels will work best to control formation of ozone.
DISCUSSION: The possible widespread use of alternative fuels in the next several years would change the emission characteristics of motor vehicles. Therefore it is important to assess the potential improvement in air quality resulting from such use. Alternative fuels are viewed as a means to improve air quality by reducing the mass emission rates from motor vehicles or by reducing the ozone-forming potential (or reactivity) of those emissions (see Chapter 12). Candidate alternative fuels include natural gas, methanol, ethanol, hydrogen, and electricity. Another candidate fuel is reformulated gasoline4, whose composition has been altered to make exhaust products less photochemically reactive and toxic and to lower total emissions of VOCs, CO, or NOx. Reformulated gasoline has the advantage that it may be used immediately in exist-
4Reformulated gasoline is not considered a true alternative fuel, but it is discussed here because its use could potentially improve air quality.
Vehicles that run on electricity or on hydrogen in fuel cells would emit virtually no precursors to ozone, although the production of the electricity or hydrogen can contribute to ozone formation depending on the feedstock and location of the generation facility. Modeling studies show that at relatively low ambient VOC/NOx ratios (in the range of 4 to 6), the use of natural gas as a motor fuel could reduce ozone formation on a mass basis (grams of ozone formed per gram of VOC emitted) by as much as 75% compared with the use of conventional gasoline vehicles. Emissions from methanol-fueled vehicles are largely methanol and formaldehyde. On a per-mass-emitted basis, methanol-fueled vehicles are predicted to reduce ozone formation by 15-40% relative to conventionally fueled vehicles if formaldehyde emissions are controlled. Ethanol would provide less benefit, especially if an increased vapor pressure is allowed. At higher VOC/NOx ratios, less improvement is expected from VOC reactivity reduction.
When considered for their effect on ozone concentrations in an urban area, alternative fuels are predicted to provide considerably less benefit to air quality than projected by the numbers cited above because motor vehicle emissions do not constitute the total of VOC emissions. For example, use of methanol in Los Angeles is predicted to lead to no more than a 10-15% reduction in ozone exposure, and little decrease in peak ozone. A major uncertainty is the effect of alternative fuels on the in-use emissions of the motor vehicle fleet, which is dominated by a relatively small fraction of vehicles having high VOC emissions. Under NOx-limited conditions, typified by high VOC/NOx ratios such as those found in Houston or Atlanta, reducing the reactivity of emissions probably would have little benefit. In high NOx regions, reducing reac-tivities would complement NOx control that might otherwise lead to local ozone increases.
Alone, no alternative fuel will solve the air pollution problems that face most large cities. Each fuel must be considered in conjunction with other controls. Specific fuels could work effectively in some regions, but would provide little benefit if used in others.
A Research Program on Tropospheric Ozone
FINDING: Progress toward reducing ozone concentrations in the United States has been severely hampered by the lack of a coordinated national research program directed at elucidating the chemical, physical, and meteorological processes that control ozone formation and concentrations over North America.
RECOMMENDATION: A coherent and focused national program should be established for the study of tropospheric ozone and related aspects of air quality in North America. This program should include coordinated field measurements, laboratory studies, and numerical modeling that will lead to a better predictive capability. In particular, the program should elucidate the response of ambient ozone concentrations to possible regulatory actions or to natural changes in atmospheric composition or climate. To avoid conflict between the long-term planning essential for scientific research and the immediacy of requirements imposed on regulatory agencies, the research program should be managed independently from the EPA office that develops regulations under the Clean Air Act and from other government offices that develop regulations. The research program must have a long-term commitment to fund research on tropospheric ozone. The direction and goals of this fundamental research program should not be subjected to short-term perturbations or other influences arising from ongoing debates over policy strategies and regulatory issues. The program should also be broadly based to draw on the best atmospheric scientists available in the nation's academic, government, industrial, and contract research laboratories. Further, the national program should foster international exchange and scientific evaluations of global tropospheric ozone and its importance in atmospheric chemistry and climate change. The recommended tropospheric ozone research program should be carefully coordinated with the Global Tropospheric Chemistry Program currently funded and coordinated by the National Science Foundation (NSF) and with corresponding global change programs in the National Aeronautics and Space Administration (NASA), the National Oceanic and Atmospheric Administration (NOAA), the Department of Energy (DoE), and other agencies.
DISCUSSION: A good analogy for the research program needed is the U.S. effort to address depletion of the stratospheric ozone layer by chloro-fluorocarbons.5 For this program, EPA is the relevant regulatory agency, but NASA's Upper Atmosphere Research Program was directed in the Clean Air Act amendments of 1977 to ''continue programs of research, technology, and monitoring of the phenomena of the stratosphere for the purpose of understanding the physics and chemistry of the stratosphere and for the early detection of potentially harmful changes in the ozone of the stratosphere.'' The partnership has worked well, and the basic research program has prepared the
5This program is discussed as an example because it has many features that would be desirable in a tropospheric ozone research program. The committee does not recommend which agency should direct such a program.
scientific foundation for international assessment and for the Montreal Protocol on Substances that Deplete the Ozone Layer (1987). NASA has developed a basic research program of laboratory and field measurements, satellite data analysis, and theoretical modeling. The particular strengths of the program have been its broad participation base, which draws on academic, government, industrial, and contract research groups, and its careful coordination with other federal and industrial programs and non-U.S. research efforts. The results of this comprehensive and coordinated research effort have been reported to Congress and to EPA. Its scientific assessments often include specific modeling studies that meet the regulatory and policy needs of EPA. A similar partnership that meets the needs of the research community and those of regulatory agencies will be necessary to establish a reliable scientific basis for the improvement of the nation's air quality.