Page 351

11
Voc Versus Nox Controls

Introduction

Although ozone concentrations in many areas of the United States violate the National Ambient Air Quality Standards (NAAQS), the circumstances of the violations can be quite varied. A few areas, such as Los Angeles, California, are isolated from regional influences, although most are not. Some have relatively high concentrations of VOCs (volatile organic compounds) compared with NOx (the oxides of nitrogen), whereas others do not. One goal of this report is to assess current understanding of the relative effectiveness of VOC versus NOx controls in ozone abatement in the United States. Knowledge of the atmospheric chemistry leading to ozone formation, together with the use of ozone isopleth diagrams (Chapter 6), provides a qualitative understanding of the relationship between ozone concentrations and VOC and NOx emissions. To actually evaluate the effectiveness of potential control strategies requires the use of photochemical air quality models (Chapter 10) that incorporate the best possible information about an area's initial and boundary conditions, emissions, and meteorology. In this chapter, we synthesize and assess much of the information from air quality models about the relative effectiveness of VOC and NOx controls in various regions of the country.

The most widely used method for determining ozone control requirements for urban areas has been the U.S. Environmental Protection Agency's Empirical Kinetic Modeling Approach (EKMA). The limitations of EKMA are discussed in Chapter 6: In practice, only periods of less than one day are simulated, and, as a result, the method cannot capture the multiday nature of



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 351
Page 351 11 Voc Versus Nox Controls Introduction Although ozone concentrations in many areas of the United States violate the National Ambient Air Quality Standards (NAAQS), the circumstances of the violations can be quite varied. A few areas, such as Los Angeles, California, are isolated from regional influences, although most are not. Some have relatively high concentrations of VOCs (volatile organic compounds) compared with NOx (the oxides of nitrogen), whereas others do not. One goal of this report is to assess current understanding of the relative effectiveness of VOC versus NOx controls in ozone abatement in the United States. Knowledge of the atmospheric chemistry leading to ozone formation, together with the use of ozone isopleth diagrams (Chapter 6), provides a qualitative understanding of the relationship between ozone concentrations and VOC and NOx emissions. To actually evaluate the effectiveness of potential control strategies requires the use of photochemical air quality models (Chapter 10) that incorporate the best possible information about an area's initial and boundary conditions, emissions, and meteorology. In this chapter, we synthesize and assess much of the information from air quality models about the relative effectiveness of VOC and NOx controls in various regions of the country. The most widely used method for determining ozone control requirements for urban areas has been the U.S. Environmental Protection Agency's Empirical Kinetic Modeling Approach (EKMA). The limitations of EKMA are discussed in Chapter 6: In practice, only periods of less than one day are simulated, and, as a result, the method cannot capture the multiday nature of

OCR for page 351
Page 352 episodes of high concentrations of ozone. EKMA also simulates ozone formation only along a single trajectory, not providing any regionwide information about the effects of controls. The number of analyses carried out using grid-based air quality models is limited but growing (see Chapter 10). Urban scale models, such as the urban airshed model (UAM) and the CIT model, have been applied to a number of cities in the United States and elsewhere. The Regional Oxidant Model (ROM) has been applied recently to the northeastern United States and to urban areas in that region. Figure 11-1 Ozone isopleth diagram for three cities (A, B, and C) that have the same peak 1-hour  ozone concentrations (Cp). The VOC/NOx ratios differ: a low ratio (c), a high ratio (B),  and a medium ratio (A). Isopleths = lines of constant 1-hour peak ozone. Ekma-Based Studies EKMA is used to generate ozone isopleth diagrams for cities, and EPA

OCR for page 351
Page 353 and other agencies have used it to determine the fractional VOC and NOx reductions needed to meet the ozone NAAQS from particular base-year conditions. The ozone isopleth diagram, introduced in Chapter 6, shows the peak 1-hr ozone concentration in terms of the initial VOC and NOx concentrations. Figure 11-1 shows a hypothetical diagram for three cities that have the same peak 1-hr ozone concentration but for which the ambient ratios of VOCs to NOx differ: a low ratio (city C), a high ratio (city B), and a medium ratio (city A). As illustrated in Figure 11-1, at low VOC/NOx (city C), reductions in NOx can have little effect or actually can cause increases in ozone; for city B, reductions in NOx can lead to substantial decreases in ozone. At moderate VOC/NOx (city A), reductions in NOx can lead to small or moderate decreases in ozone, depending on the shape of the isopleth and the amount of NOx reduction. City A is located along what is often called the ridge line. If the molar ratio of carbon to NOx is greater than about 20 ppbC/ppb, NOx control is clearly more effective than VOC control, whereas at a ratio of about 10 or less, VOC control is more effective. At ratios between 10 and 20, control of either VOC or NOx or both might be preferred; specific situations must be carefully evaluated to determine the relative effectiveness of alternative abatement strategies (Blanchard et al., 1991). Although the VOC/NOx ratio is a useful measure of the overall nature of the VOC-NOx-ozone system, it is at best a qualitative measure of the reactivity of a given city's air because, as noted in Chapter 6, VOC/NOx ratios vary both spatially and diurnally in a given city and from one episode to the next for the same city. Variation among proximate cities is observed as one travels from west to east in the Los Angeles basin; the VOC/NOx ratio in the atmosphere varies from that of city C to that of city A and, as one goes sufficiently far east, to that of city B. In areas where the VOC/NOx ratio is between roughly 10 and 20, control of NOx may reduce the effectiveness of VOC controls. At a fixed level of VOC emissions, NOx control in such cases may cause ozone concentrations to decrease in downwind areas and increase in near-source areas. In some downwind areas, ozone concentrations may decrease less than they would have decreased if VOC emissions alone had been reduced (Blanchard et al., 1991). Results of sample State Implementation Plans (SIPs) for various cities generated by EKMA are given in Table 11-1. In each case, only VOC control was considered. In general, the higher the original ozone concentration, the greater the VOC control predicted. (Biogenic VOC emissions are not accounted for in the calculations in Table 11-1.) Chang et al. (1989) used EKMA to study the effect of conventional and methanol-fueled vehicles on air quality in 20 cities. In that study they calculated the effect that removing light-duty VOC emissions (primarily emissions

OCR for page 351
Page 354 TABLE 11-1 Ozone Design Values, VOC Concentrations, VOC/NOx, Mobile Source Emissions, and Estimated VOC control requirementsa,b Area Ozone design value, ppb Median VOC, ppbC Median VOC/NOx On-road mobile source percent of emissionse Required VOC % control to meet standardd   VOC NOx Akron, Ohio 125 600 12.8 39 —e — Atlanta, Georgia 166 600 10.4 52 43 25 - 50 Boston, Massachusetts 165 380 7.6 51 47 35 Charlotte, North Carolina 149 390 10.4 52 — 25 - 50 Cincinnati, Ohio 157 740 9.1 42 40 > 50 (Table continued on next page)

OCR for page 351
Page 355 (Table continued from previous page) Area Ozone design value, ppb Median VOC, ppbC Median VOC/NOx On-road mobile source percent of emissionse Required VOC % control to meet standardd   VOC NOx Cleveland, Ohio 145 780 7.5 49 36 25 50 Dallas, Texas 160 730 11.8 52 43 25 50 El Paso, Texas 160 670 11.9 66 — 25 50 Fort Worth, Texas 160 630 11.8 52 — — Houston, Texas 200 740 12.9 36 — > 50 Indianapolis, Indiana 130 690 10.9 49 — — Kansas City, Missouri 130 410 8.5 50 — — Los Angeles, California 360 — 7.8 46 60 85 Memphis, Tennessee 146 127 13.9 48 — 25 50 Miami, Florida 130 103 13.3 61 — 25 50 New York, New York 217 — 9.6 47 40 — Philadelphia, Pennsylvania 180 570 8.0 41 41 25 50 Portland, Maine 140 430 11.6 38 — 0 15 Richmond, Virginia 125 450 11.2 44 — 25 50 St. Louis, Missouri 160 570 9.6 43 23 25 50 Washington, D.C. 140 600 8.7 55 56 25 50 (Table continued on next page)

OCR for page 351
Page 356 (Table continued from previous page) Area Ozone design value, ppb Median VOC, ppbC Median VOC/NOx On-road mobile source percent of emissionse Required VOC % control to meet standardd   VOC NOx Wilkes Barre, Pennsylvania 125 430 14.3 43 — — Average of nonattainment areas — — — 48 50 — aBased on Chang et al. (1989), SCAQMD (1989), OTA (1989), EPA (1983), Systems Applications Incorporated (1990), E.H. Pechan (1990). b''VOC'' refers to volatile organic compounds excluding methane. cMobile-source percentage of total NOx emissions primarily derived from EPA (1983) and might not correspond to same period used to derive mobile-source VOC percentage from Chang et al. (1989). dBased on EKMA. A range indicates uncertainty in the amount of control needed. e—No data were available or no calculation was made.

OCR for page 351
Page 357 from automobiles and pickup trucks) would have on ozone concentrations, and how varying NOx emissions would affect these results. The response—the percentage of ozone reduction achieved in relation to a given percentage VOC reduction—is highly variable. The ratio of those two can be construed as the sensitivity of ozone formation to VOC emissions, and is a measure of the effectiveness of VOC control. The sensitivity of ozone to VOCs for 20 cities is given in Table 11-2. The sensitivity is seen to correlate with the ambient VOC/NOx ratio in that higher sensitivities are associated with lower ratios. Biogenic emissions were not included in the calculation. At VOC/NOx ratios less than 8, reductions in VOCs were found to be particularly effective, and at higher ratios, the sensitivity slowly declines as the ratio increases. Carter and Atkinson (1989b) obtained similar results for air parcels in urban areas. Most assessments of control strategies using EKMA have not considered the effects of biogenic VOCs. Chameides et al. (1988) argued that biogenic VOCs must be considered, particularly in southern cities where warm temperatures lead to significant emissions of isoprene. They showed that for Atlanta, anthropogenic VOCs need to be reduced by 30% to meet the NAAQS according to the standard EKMA calculation with no biogenic VOCs. Inclusion of isoprene increases the necessary reduction in anthropogenic VOCs to 70%. With inclusion of other biogenic hydrocarbons, ozone concentrations were predicted to exceed the NAAQS with no anthropogenic VOC emissions. Once isoprene is included, the percent reduction in NOx emissions needed to meet the NAAQS is less than the required reduction in VOC emissions. Intercomparison of model results indicates that EKMA and other single-and double-layer trajectory models are too limited by their mathematical formulation and lack of physical detail to assess ozone control strategies accurately. A major, shortcoming is that high-ozone episodes are multiday events, and EKMA simulations are generally less than one day long. NOx is removed from the photochemical system faster than the bulk of the VOCs, leading to more NOx-limited conditions on subsequent days of an episode. Grid-based airshed models provide a much stronger foundation on which to build ozone control strategies. Because of EKMA's inherent limitations, our assessment of the relative effectiveness of VOC and NOx controls will focus on applications of grid-based models.

OCR for page 351
Page 358 TABLE 11-2 Sensitivity of Ozone Formation to VOC Emissions Area Median VOC/NOx Sensitivity to light-duty vehicle VOC emissions, D O3/DVOCa Akron, Ohio 12.8 0.44 Atlanta, Georgia 10.4 0.56 Boston, Massachusetts 7.6 1.08 Charlotte, North Carolina 10.4 0.54 Cincinnati, Ohio 9.1 0.52 Cleveland, Ohio 7.5 0.92 Dallas, Texas 11.8 0.53 El Paso, Texas 11.9 0.54 Fort Worth, Texas 11.8 0.51 Houston, Texas 12.9 0.59 Indianapolis, Indiana 10.9 0.51 Kansas City, Missouri 8.5 0.67 Memphis, Tennessee 13.9 0.45 Miami, Florida 13.3 0.55 Philadelphia, Pennsylvania 8.0 1.45 Portland, Maine 11.6 0.43 Richmond, Virginia 11.2 0.49 St. Louis, Missouri 9.6 0.58 Washington, D.C. 8.7 0.64 Wilkes Barre, Pennsylvania 14.3 0.44 Average   0.62 aD O3/D VOC, ratio of percent reduction in ozone concentration to percent reduction in VOC emissions. Source: Chang et al., 1989.

OCR for page 351
Page 359 Grid-Based Modeling Studies Two areas of the United States—the Northeast corridor, which extends from the Washington, D.C. area to beyond Boston, and the Los Angeles basin—have received a large share of attention in the evaluation of ozone abatement strategies. Los Angeles has little influence from upwind sources, whereas each city in the Northeast corridor is affected by transport of ozone and precursors from upwind regions. In essence, the Northeast corridor acts as a system, and the effectiveness of ozone controls in one urban location will depend on controls throughout the region. Limited studies are available for other areas of the country, such as the Southeast and the Midwest. Los Angeles Basin The effects of controlling VOC and NOx emissions in the Los Angeles basin have been explored in a variety of studies, for example those of the South Coast Air Quality Management District (1989) and Milford et al. (1989). Basinwide control of VOC emissions was predicted in these reports to reduce ozone concentrations everywhere. Controlling NOx emissions was predicted to lead to increased ozone concentrations in the downtown and midbasin areas but decreased ozone concentrations in the far eastern portion of the region. Since the emissions inventories used in those studies apparently underpredicted VOCs, it is likely that a larger portion of the basin would respond favorably to NOx reduction than was predicted. Studies of ozone abatement strategies in Los Angeles have used both the urban airshed model (UAM) and the CIT model (see Table 10-1). The UAM was used in developing the air quality management plan for the South Coast air basin (SCAQMD, 1989). The CIT model was used to determine the effects of VOC and NOx controls on ozone, nitric acid (HNO3), nitrogen dioxide (NO2), peroxyacetyl nitrate (PAN), and aerosol nitrate for an episode that occurred Aug. 30-31, 1982 (Russell et al., 1988a,b, 1989; Milford et al., 1989). Milford et al. (1989) used the CIT Model to develop ozone isopleth diagrams across the Los Angeles Basin, showing how the effectiveness of NOx and VOC controls varies spatially (Figure 11-2). Likewise, they developed the isopleth diagram for peak ozone in the basin (Figure 6-4). In those figures, the base level of emissions corresponds to the upper right hand corner, and increasing levels of VOC and NOx control are plotted along the horizontal and vertical axes, respectively. Milford et al. (1989) found that NOx controls were most effective in the downwind regions, e.g., around San Bernardino.

OCR for page 351
Page 360 Figure 11-2 Ozone isopleths for locations within the Los Angeles air basin from an airshed  model for spatially uniform reductions of VOC and NOx. Source: Milford et al., 1989. These regions also had the highest ozone concentrations. In the central regions, such as downtown Los Angeles and Pasadena, where peak ozone concentrations were lower, VOC controls were most effective, and NOx reduc-

OCR for page 351
Page 361 tions could inhibit ozone reduction. Control of both VOCs and NOx led to basinwide reductions. Correction of the likely underestimation of VOC and CO emissions from motor vehicles would enhance the effectiveness of NOx controls. The spatial variation in ozone response has been explained by Milford et al. (1989) based on the emission patterns. In Los Angeles and Pasadena, the VOC/NOx ratio is estimated to be between 5 and 10 at 9:00 a.m. Especially in Pasadena, this ratio does not increase dramatically by noon, apparently because of high local NOx emissions compared to downwind locations. Thus the situation corresponds to the region to the left of the ridge line in Figure 11-1, and NOx controls can lead to increased ozone. At the downwind locations, however, there are lower local emissions, and much of the NOx has been lost due to deposition and chemical transformations. The resulting VOC/NOx ratio is much larger , corresponding to the region to the right of the ridge line in Figure 11-1, where the chemistry is NOx limited, and hence NOx controls are most effective. As noted in Chapter 6, a significant advantage of using grid-based airshed models to generate ozone isopleths is that these models show the effect of precursor controls on peak concentrations of ozone, regardless of where the peak occurs in the air basin. For example, in the studies of Milford et al. (1989), reducing NOx or VOCs by 25%, or each by 15%, shifts the location of the ozone peak westward from San Bernardino to Chino. Figure 11-3 shows the effect of VOC and NOx controls on peak O3 in the Los Angeles air basin as a whole, i.e. irrespective of the location of peak ozone. A more L-shaped isopleth results. The isopleths in Figure 11-3 show that when the Los Angeles air basin as a whole is considered, up to 80% control of VOCs alone will not result in attainment of the NAAQS. Tesche and McNally (1990) applied the UAM to the South Central Coast air basin, in the Santa Barbara area of California, for Sept. 5-7, and Sept. 16-17, 1984. They predicted ozone isopleths for this air basin which, like those of Milford et al. (1989) for Los Angeles, are more L-shaped than are the EKMA-type isopleths shown in Figure 11-1. As Milford et al. (1989) and Tesche and McNally (1990) pointed out, although the calculations are specific to the southern California area, the approach and issues involved (e.g. downwind areas) have general validity and applicability. Nonlinearities in the response of ozone concentrations to emissions changes generally result in smaller ozone reductions than might be expected or desired from reducing emissions. For example, by the year 2000, mobile sources in Los Angeles are expected to account for about 30% of total VOC emissions. Airshed model calculations indicate that removing this fraction of VOCs would decrease peak ozone 16% from 270 to 230 ppb for the particular set of

OCR for page 351
Page 368 (Table continued from previous page) Episode Base case or control strategy Model Purposea Seasonal extrapolationb August 22-31, 1980 Base case ROM2.0 R     VOC1 ROM2.0 R     VOC2 ROM2.0 R     NOxl ROM2.0 R     NOx2 ROM2.0 R     VOC2 + ROM2.0 R     NOx3 ROM2.0 R     VOC3 + ROM2.0 R,S     NOx3         With TSDF emissions       Southeastern U.S. Domain: April 14-29, 1980 Base case ROM2.0 R     VOC1 ROM2.0 R     VOC2 ROM2.0 R     VOC2 + ROM2.0 R     NOx3       June 29-July 14, 1980 Base case ROM2.0 R     VOC1 ROM2.0 R     VOC2 ROM2.0 R     VOC2 + ROM2.0 R     NOx3       August 10-September 1, 1980 Base case ROM2.0 R     VOC1 ROM2.0 R     VOC2 ROM2.0 R     VOC2 + ROM2.0 R     NOx3 With TSDF emissions   R,S   aE, Model evaluation study, S, model sensitivity analysis, R., regulatory analysis. b, This simulation was used, along with others, to extrapolate episodic model results to a full season. cTSDF, treatment, storage, and disposal facility.

OCR for page 351
Page 369 Figure 11-4a Predicted episode maximum ozone concentrations (ppb) for  the 1985 base case (July 2-17, 1988). Source: Possiel et al., 1990. Possiel et al. (1990) examined the effects of proposed regional control strategies on ozone in the Northeast for July 2-17, 1988, the most severe episode in this region between 1980 and 1988. The model was run for a base case of 1985. Its emissions data were taken from the 1985 NAPAP inventory, adapted for the above-average temperatures that prevailed during the episode, and from Pierce et al. (1990) for biogenic emissions. The target year was 2005, with projected emissions that accounted for existing federal and state controls. With these controls, anthropogenic VOC emissions were 20% lower than in 1985, carbon monoxide emissions were 43% lower, and total NOx emissions were the same. Several other control scenarios were also applied to the 2005 calculation. Results for the 1985 base case and for the 2005 case with existing controls are shown in Figure 11-4a,b in terms of maximum ozone concentrations for the episode. Predicted reductions in peak concentrations ranged from 5-10% in and downwind of most major source areas, to as much as 20% in New York City. For both simulations, ozone concentrations exceeded 120 ppb r and downwind of all major source regions. Changes in ozone relative to the 2005 case with existing controls are shown in Figure 11-5 for two scenarios: In one scenario (Figure 11-5a), VOC emissions were reduced throughout the United States, leading to a 45% reduction; other emissions were at levels

OCR for page 351
Page 370 Figure 11-4b Predicted episode maximum ozone concentrations (ppb) for the 2005 case  with existing controls (July 2-17, 1988). Source: Possiel et al., 1990. assuming existing controls. In the other scenario (Figure 11-5b), anthropogenic VOCs were reduced by 49% in the Northeast corridor and by 26% elsewhere, with corresponding reductions in NOx of 26% and 34%, respectively. The reduction in VOCs alone produced the greatest effect north of Philadelphia, with reductions in peak ozone of as much as 25-50% in the immediate area of New York City. There was little change elsewhere, including most of New England and the southern part of the corridor. The predictions for the combined NOx-VOC strategy were quite different. Peak ozone was reduced by 10-15% across much of the domain, and the reductions were generally greater than with the VOC-only strategy. In the New York City area, however, the combined controls were less effective in reducing ozone. The frequency distribution of maximum ozone concentrations was also examined for the two scenarios. The combined strategy was more effective in the Washington-Baltimore area, Philadelphia, and Boston; the VOC-only strategy was more effective in New York City. There was little difference between the two strategies in Connecticut. The combined strategy led to decreased population exposure in all regions of the corridor except the New York City area, but 43% of the corridor's population lives there. Even with the control strategies, ozone concentrations were predicted to exceed 120 ppb

OCR for page 351
Page 371 Figure 11-5a Percentage change in episode maximum ozone concentrations, 2005 base  case versus a VOC-alone reduction strategy (July 2-17, 1988). along the Northeast corridor. Possiel and Cox (1990) examined another set of control scenarios for the period July 2-17, 1988; they showed results for NOx control alone, VOC control alone, and simultaneous NOx and VOC control, all relative to a somewhat different base case for 2005 emissions (see Table 11-4). These scenarios, based on application of ''maximum technology,'' resulted in reduction of NOx by 58% and anthropogenic VOCs by 63% compared with the 2005 base case. Total VOC emissions were reduced by 40% within the Northeast corridor but by only 20% outside the corridor, because of the preponderance of natural emissions there. Control of NOx alone caused large reductions in ozone throughout most of the U.S. portion of the model domain, including the Northeast corridor. The exception was New York City, where ozone increased (see Table 11-5). Outside the Northeast corridor, control of VOCs alone led to ozone concentrations 15-25 ppb higher than did control of NOx alone. However, the VOC-only strategy was more effective in lowering peak ozone in New York City. VOC control resulted in a much larger area with ozone above 120 ppb than did NOx control. Combined control of NOx and VOCs reduced ozone outside the Northeast corridor only slightly more (< 5%) than did NOx control alone. Within the corridor, the combined controls were

OCR for page 351
Page 372 Figure 11-5b Percentage change in episode maximum ozone concentrations, 2005 base case versus a combined NOx-VOC reduction strategy (July 2-17, 1988). Source: Possiel et al., 1990. more effective than either alone, except in New York City, where VOC-only control was most effective. McKeen et al. (1991b) also found that control of NOx was more effective than control of VOCs in reducing peak ozone values across most of the eastern United States and that control of NOx alone led to increases in ozone in a few areas of high NOx emissions. Their model had a grid size of 60 km and simulated a different meteorological period. Nevertheless, this model gave results similar to those of ROM for similar scenarios. The results from ROM for the effect of a NOx-VOC versus VOC-only strategy agree with analyses based on much simpler models (Sillman et al., 1990a,b), as do the studies of the role of biogenic VOCs (Trainer et al., 1987; Chameides et al., 1988; McKeen et al., 1991b). The results from ROM emphasize that a combined NOx-VOC strategy should be more effective than a VOC-only strategy in reducing ozone over a large geographic area in the Northeast. A VOC-only strategy would be more effective in some areas of high population density (New York City) but less beneficial downwind. Although the general nature of these results is likely to be correct, the details of the predictions should be viewed with caution, because of known deficiencies in the base-case emissions inventories (see Chapter 9) and because of the

OCR for page 351
Page 373 possible deficiencies in the model itself, as implied by evaluation studies (Schere and Wayland, 1989). New York Metropolitan Area Rao and co-workers (Rao et al., 1989; Rao and Sistla, 1990) have applied the UAM to study the control of ozone in the New York City area. Rao and Sistla (1990) studied imposition of 75% control on NOx, VOCs, or both (Table 11-6). Control of NOx alone decreased ozone concentrations in Connecticut and New Jersey but increased them in New York. Ozone exposure above 120 ppb increased. VOC-only control led to substantial reductions in ozone in all three regions and to a concomitant reduction in exposure, and combined controls were less effective in reducing ozone than were VOC controls alone. These results indicate that VOC controls are necessary to reduce ozone concentrations in the New York area, although downwind areas can benefit from NOx reductions. It also was found that biogenic VOC emissions alone, in concert with emissions of anthropogenic NOx, would lead to ozone concentrations above 120 ppb. The sensitivity of the peak ozone and exposures can be estimated from these results and are given in Table 11-6. The sensitivities to VOCs are for the anthropogenic portion only and are generally less than those found by Chang et al. (1989) for other cities using EKMA. The results of Chang et al. (1989) also showed that the effects of NOx and VOC controls are not additive. In accordance with the UAM study by Rao and Sistla (1990), peak ozone concentrations as predicted by ROM fell in New York in response to VOC controls but not NOx controls (Possiel and Cox, 1990). Virtually all other cities in the ROM domain responded favorably to control of VOCs or NOx or both. ROM results indicate that eight-hour exposure to ozone would decrease in all cities, including New York, when NOx controls are imposed in addition to VOC controls. This is contradictory to the UAM results discussed above. NOx reductions led to a regionwide decrease in ozone exposure. A shortcoming of current regional air quality models is that they do not have the spatial resolution required to accurately assess the chemical transformation and transport within urban areas. A solution to this problem is to embed, or "nest" a model with finer spatial resolution. For example, an urban model can be embedded in a regional model. The regional model then prescribes transport into and boundary conditions for the urban domain. A UAM-ROM interface has been developed to serve this purpose (Rao et al., 1989). In this case, the nesting is one way; information flows from ROM to UAM, but not back. The results of the nested-grid study are interesting in that they compare

OCR for page 351
Page 374 TABLE 11-5 Ozone Response in Northeast to VOC and NOx Controls Found Using ROMab Area Basec.d 63% VOC controlc 58% NOx controlc Combined VOC NOx controlc Sensitivity of peak ozone to VOC Sensitivity to NOx Washington,D.C. 140/87 132/74 122/77 113/74 0.09 0.22 Philadelphia, Pennsylvania 143/92 129/85 116/74 111/73 0.16 0.32 New York, New York 234/107 140/87 257/107 163/83 0.63 -0.17 Rhode Island 138/85 120/78 112/68 105/66 0.20 0.32 Boston, Massachusetts 137/84 121/80 103/67 100/67 0.18 0.43 Pittsburgh, Pennsylvania 136/80 126/77 103/66 101/66 0.12 0.41 Detroit, Michigan 120/76 112/74 104/70 102/69 0.11 0.22 aIn parts per billion. bFrom Possiel and Cox (1991). cFirst value is 90th percentile 1-hr ozone, and second value is peak 8-hr average. dBase case refers to the year 2005.

OCR for page 351
Page 375 TABLE 11-6 Effect of Controls on Ozone in New Yorka Peak Ozone (second day), ppb Exposure to ozone > 120 ppb (1000 population hours)   New York Connecticut   Base Case 167 173 28,127 75% VOC control 144(0.18)b 136(0.28)b 10,418(0.84) 75% NOx control 185(-0.14)b 154(0.14)b 33,120(-0.23) 75% VOC and NOx control 149 138 20,598 aDerived from Rao and Sistla (1990). bSensitivity of peak ozone and ozone exposure shown in parentheses. model calculations of a regular (nonnested) simulation with various nesting procedures (Table 11-7). Predicted ozone concentrations vary considerably depending on the kind of nesting employed. Although Rao et al. (1989) did not test how these variations affected estimates of necessary control levels, the scatter in the predictions indicates that the calculated effect of controls would differ greatly depending on which nesting procedure is used. Summary Application of grid-based air quality models to various cities and regions in the United States shows that the relative effectiveness of controls of volatile organic compounds (VOCs) and oxides of nitrogen (NOx) in ozone abatement varies widely. Most major cities experience ozone concentrations that exceed the National Ambient Air Quality Standard (NAAQS) one-hour concentration of 120 ppb—a result of the density of precursor emissions in those areas. The predominant sources of emissions are mobile, although other sources contribute significantly. These cities share an ozone problem, but differ widely in the

OCR for page 351
Page 376 TABLE 11-7 Comparison of Nesting Techniques for Peak Ozone Predictionsa, ppb   New Jersey New York Connecticut Observed 145 240 303 Regular UAMb 180 199 202 Nested, ROMc ICs, BCs, winds 155 138 148 Nested, ROM ICs, BCs 156 143 167 Nested, ROM BCs 126 259 233 aFrom Rao et al. (1989), for July 21, 1980. bUAM, urban airshed model, without nesting. cROM, Regional Oxidant Model, and refers to the ROM supplying initial conditions (ICs), boundary conditions (BCs), wind fields. magnitude of the problem and in the relative contributions of anthropogenic VOCs and NOx and biogenic emissions. As a result, the optimal set of controls relying on VOCs, NOx, or, most likely, reductions of both, will vary from one place to the next. In cities where the VOC/NOx ratio is high, VOC control provides less ozone reduction per unit of VOC reduction than in cities with a low VOC/NOx ratio. Cities with a high VOC/NOx ratio benefit from NOx control, but less so if the ratio is low. Studies have predicted that in some areas—downtown Los Angeles and New York City, for example—ozone will increase in certain locations (not necessarily those where the peak occurs), if NOx emissions are lowered. Few urban areas in the United States can be treated as isolated cities unaffected by regional sources of ozone. The regional nature of the ozone problem east of the Mississippi, as demonstrated by ozone observations (see Chapter 4) and model simulations (e.g., Roselle et al., 1990); McKeen et al., 1991a) requires the use of regional models for assessment of control strategies. The Regional Oxidant Model (ROM) has been the major tool used for regulatory studies of areas affected by regional ozone. A significantly greater effort needs to be devoted both to understanding the reason's for the model's failures and to further developing the model itself. The present regional

OCR for page 351
Page 377 models do not have sufficient spatial resolution for detailed studies of major urban areas, and a nested model approach is likely to be necessary. A fully interactive, two-way nested multiscale model is desirable for studying intercity and regional pollutant transport. Ozone air quality is intimately related to other air-quality issues, such as acid deposition and visibility, and a comprehensive modeling system with high spatial resolution is ultimately necessary. Biogenic VOCs, in combination with anthropogenic NOx, are capable of generating ozone concentrations above 80 ppb in favorable meteorological conditions across much of the eastern United States, with values of more than 100 ppb downwind of a number of major cities. Future assessments of control strategies must include biogenic emissions, given their potential for generating ozone concentrations close to the (NAAQS) concentration. Many simulations conducted to date have relied on emissions inventories that are suspected of significantly underestimating anthropogenic VOC emissions (see Chapter 9) and that have not included biogenic emissions. The result is an overestimate of the effectiveness of VOC controls and an underestimate of the efficacy of NOx controls (Chameides et al., 1988; McKeen et al., 1991b). Underestimates in the VOC inventories might be partly responsible for the underprediction of ozone concentrations in central urban areas (SCAQMD, 1989; Rao et al., 1989). The consequences of an underestimate in the VOC inventories on predicted concentrations of ozone and its precursors and on control strategies must be investigated. Even with the limitations of present models and emissions inventories, certain robust conclusions emerge when the modeling studies are synthesized. Production of ozone is limited by the availability of NOx and is much less sensitive to anthropogenic VOCs in most rural environments in the eastern United States, where NOx concentrations are less than ˜2 ppb and the VOC/ NOx ratio is high. Control of NOx is also effective in lowering peak ozone concentrations in many urban areas, although it is predicted to lead to an increase in ozone in some places, such as downtown Los Angeles and New York City. The ozone increases in these urban cores, however, are predicted to be accompanied by decreases in ozone downwind, in the Los Angeles basin and Connecticut, respectively. While control of VOCs never leads to a significant increase in ozone, there are many areas where control of VOCs is either ineffective or does not bring an area into compliance with the NAAQS. Hence NOx control will probably be necessary in addition to or instead of VOC control to alleviate the ozone problem in many cities and regions. The optimal set of controls of NOx, VOCs, or both will vary from one region to another, as discussed above.

OCR for page 351