Modeling the Oxidative Capacity of the Atmosphere of the South Coast

Department of Earth Sciences, University of New Hampshire,. Durham, New Hampshire 03824. The production of HOx radicals in the South Coast Air. Basin ...
0 downloads 0 Views 100KB Size
Environ. Sci. Technol. 2004, 38, 753-757

Modeling the Oxidative Capacity of the Atmosphere of the South Coast Air Basin of California. 2. HOx Radical Production ROBERT J. GRIFFIN† Department of Civil and Environmental Engineering, Duke University, Durham, North Carolina 27708, and Institute for the Study of Earth, Oceans, and Space and Department of Earth Sciences, University of New Hampshire, Durham, New Hampshire 03824

The production of HOx radicals in the South Coast Air Basin of California is investigated during the smog episode of September 9, 1993 using the California Institute of Technology (CIT) air-quality model. Sources of HOx (hydroxyl, hydroperoxy, and organic peroxy radicals) incorporated into the associated gas-phase chemical mechanism include the combination of excited-state singlet oxygen (formed from ozone (O3) photolysis (hν)) with water, the photolysis of nitrous acid, hydrogen peroxide (H2O2), and carbonyl compounds (formaldehyde (HCHO) or higher aldehydes and ketones), the consumption of aldehydes and alkenes (ALK) by the nitrate radical, and the consumption of alkenes by O3 and the oxygen atom (O). At a given time or location for surface cells and vertical averages, each route of HOx formation may be the greatest contributor to overall formation except HCHO-hν, H2O2-hν, and ALKO, the latter two of which are insignificant pathways in general. The contribution of the ALK-O3 pathway is dependent on the stoichiometric yield of OH, but this pathway, at least for the studied smog episode, may not be as generally significant as previous research suggests. Future emissions scenarios yield lower total HOx production rates and a shift in the relative importance of individual pathways.

Introduction Hydroxyl (OH) and peroxy (hydroperoxy (HO2) and organic peroxy (RO2)) radicals play key roles in atmospheric chemistry. OH is the primary oxidant that consumes carbon monoxide (CO) and volatile organic compounds (VOCs), leading to the formation of HO2 and RO2. Peroxy radicals convert nitrogen oxide (NO) to nitrogen dioxide (NO2), the photolysis of which leads to tropospheric ozone (O3) production (1). In this study, OH, HO2, and RO2 are together defined as HOx. HOx forms via numerous atmospheric chemical pathways. The photolysis of O3 results in molecular oxygen (O2) and either oxygen atom (O) or excited-state singlet oxygen (O(1D)), the latter of which reacts with water (H2O) to form two molecules of OH (1): † Phone: (603) 862-2021; Fax: (603) 862-2124; E-mail: rjg@ gust.sr.unh.edu.

10.1021/es034129v CCC: $27.50 Published on Web 12/23/2003

 2004 American Chemical Society

O3 + hν f O2 + O(1D) 1

O( D) + H2O f 2OH

(1a) (1b)

where hν represents light at the appropriate photolytic wavelength. Photolysis of nitrous acid (HONO), hydrogen peroxide (H2O2), formaldehyde (HCHO), and higher carbonyls (CARB) leads to HOx production (1-3):

HONO + hν f 0.9NO + 0.9OH + 0.1NO2 + 0.1HO2 (2a) H2O2 + hν f 2OH

(2b)

HCHO + hν f CO + 2HO2

(2c)

RC(O)H + hν f RC(O)O2 + HO2

(2d)

RC(O)R + hν f RC(O)O2 + RO2

(2e)

where R represents an alkyl group, RC(O)H represents higher aldehydes, RC(O)O2 is an acyl peroxy radical, and RC(O)R represents a ketone. In reactions 2a and 2c-e, O2 is assumed to add instantaneously to hydrogen atoms (H) to form HO2 and alkyl or acyl radicals to form RO2 or RC(O)O2, respectively. Yields shown in reaction 2a are based on the SAPRC-99 mechanism (4). It could be argued that the photolysis of HONO should not be included because it may have been formed from the combination of OH and NO very shortly before it photolyzes. However, the combination of OH and NO is only a minor route of HONO formation. Therefore, its photolysis is included as a HOx source. Oxidation of alkenes (ALK) and aldehydes (ALD) by the nitrate radical (NO3) contributes to HOx production (1-3, 5):

R1R2dR3R4 + NO3 f R1R2(ONO2)R3R4O2

(3a)

RC(O)H + NO3 f RC(O)O2 + HNO3

(3b)

where R1R2dR3R4 represents an alkene. Oxidation of alkenes by NO3 proceeds via addition to the double bond (1-3, 5). Oxidation of aldehydes is initiated by NO3 abstracting the aldehydic H to form nitric acid (HNO3) (2, 3). Both are followed rapidly by O2 addition. NO3 is relevant only at night because its daytime concentrations are extremely small (1). Consumption of alkenes by O3 and O produces HOx:

where the dot indicates a radical. O3-initiated oxidation occurs via the bridging of the double bond. The resulting ozonide decomposes with the product distribution depending upon the structure of the parent (1-3, 5). Alkene-oxygen atom reactions proceed via addition to one end of the double bond VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

753

followed by collisional stabilization, forming carbonyls or epoxides, or decomposition, resulting in HOx. Again, the product distribution depends on the structure of the parent alkene (5). This study is the first to apply a three-dimensional episodic air-quality model including spatially and temporally resolved emissions and meteorology to study HOx production from individual pathways. Other relevant studies have investigated HOx production rates using a box model, while focusing on the continental boundary layer in fall and winter (6), Los Angeles (7), and a rural atmosphere (7). Paulson and Orlando (7) indicate that the reaction of alkenes with O3 is the dominant mechanism of HOx formation in average surfacelevel urban polluted air and that the combination of O(1D) with H2O is dominant in rural areas. The relative contribution of the alkene-ozone (ALK-O3) pathway increases at night. Ariya et al. (6) also indicate that the ALK-O3 pathway is especially important at night and during seasons with shorter days. While not focusing on individual pathways, numerous studies indicate the importance of accurate observation and modeling of HOx concentrations. Most attempt to match simulated concentrations to ambient measurements using a simplified photochemical box or one-dimensional model. Such studies have focused on the marine boundary layer (8-13), urban areas (14-16), or rural continental regions (14, 17). One study focused on nighttime concentrations of OH using a global chemical transport model (18). Except in isolated instances, these studies show that many kinetic photochemical models overpredict HOx concentrations up to 50% compared to observations. In most cases, the photochemical models are run with observed ambient concentrations as initial conditions or constraints; phenomena such as advection, changes in meteorology, and emissions are not always considered. As part of the Los Angeles Free Radical Experiment, measurements of OH and HO2 were made September 24 through September 26, 1993 in Claremont, a downwind (dominant wind patterns are west to east) suburb of Central Los Angeles (CELA) (16). Although these dates are slightly later than the date modeled for the current work, as discussed in the following section, the observations give a qualitative indication of the levels of HOx in the polluted South Coast Air Basin (SoCAB) of California atmosphere on a late summer day. Mixing ratios of both OH and HO2 peak during midday, coincident with times of strongest photochemistry. While observations of OH show a maximum of approximately 6 × 106 molecules cm-3 close to 1200 h on both September 24 and 25, 1993, the observations of HO2 show a maximum of approximately 2 × 108 close to 1500 h on September 25, 1993 (16).

Methodology The California Institute of Technology (CIT) three-dimensional atmospheric chemical transport model is used to investigate the spatial and temporal production of HOx in the SoCAB on September 9, 1993. A description of this model is given in part I, as is information pertaining to meteorology, emissions, and boundary and initial conditions (19). The CIT model uses this information in conjunction with the Caltech Atmospheric Chemistry Mechanism (CACM) (20) to solve the atmospheric convective diffusion equation for temporal and spatial concentration profiles. Model calculations are initiated on the afternoon of September 7, 1993 to minimize the effect of any bias in the initial conditions. Hourly average concentrations predicted using the CIT model and rate constants and stoichiometric yields implemented in CACM (20) are utilized to calculate HOx production rates: 754

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 3, 2004

O(1D)H2O ) 2.0k1b[O(1D)][H2O]

(5a)

HONOhν ) k2ahν[HONO]

(5b)

H2O2(hν) ) 2.0k2bhν[H2O2]

(5c)

HCHOhν ) 2.0k2chν[HCHO]

(5d)

CARBhν ) 2.0hν

∑k

CARBhν,i[CARB]i

(5e)

i

ALKNO3 ) [NO3]

ALK-NO3,i[ALK]i

(5f)

∑k

ALD-NO3,i[ALD]i

(5g)

∑χ k

(5h)

∑δ k

(5i)

ALDNO3 ) [NO3] ALKO3 ) [O3] ALKO ) [O]

∑k

i

i

i

i ALK-O3,i[ALK]i

i ALK-O,i[ALK]i

i

where production rates of HOx (ppm min-1) are shown with the main species subscripted by the consuming species or phenomenon (e.g., HCHOhν ) formation rate through formaldehyde photolysis), kj represents either photolysis (solar zenith angle dependent, min-1) or bimolecular (ppm-1 min-1) kinetic rate constants (numerical subscripts given in text), hν represents a meteorological parameter describing solar intensity (unity indicates clear sky conditions at noon at the top of the model domain), χi and δi represent molar stoichiometric yields of HOx in the given reactions, and the bracket notation represents mixing ratios (ppm). Spatially and temporally resolved humidity fields determine [H2O]. Rates are converted to million molecules cm-3 s-1 assuming a constant temperature of 300 K.

Results and Discussion Base Case. While only three locations cannot serve to characterize the entire SoCAB, Figures 1-3 show simulated HOx formation rates for CELA, Azusa (AZUS), and Riverside (RIVR), respectively, on September 9, 1993. These three locations are illustrative of the general character of the SoCAB atmosphere on this date. Primary emissions dominate in CELA due to proximity to the hub of the region’s freeway system. AZUS and RIVR are downwind locations typified by secondary photochemical activity and transport of pollutants from upwind. See Part I in this series for a map of the SoCAB (19). Production rates vertically averaged (weighted by the size of the model layer) over the five-layer, 1.1-km model domain are indicated in Figures 1-3. During the day, surface values are higher (up to 45%) than those that are vertically averaged; the contrary is true at night when NO3 reacts with alkenes and aldehydes that have been mixed vertically. However, at night total HOx production is generally small compared to the daytime. Figures 1-3 indicate both the temporal and spatial dependence of simulated HOx formation rates. Total production rates are greatest in downwind locations and exhibit maximum peaks in early afternoon in each location, corresponding to peak photochemical activity; other local maxima in the total production rates correspond to sunset when photolysis ceases and high concentrations of NO2 and O3 combine to form NO3. Instantaneous peak and 24-h average production rates for September 9, 1993 for each pathway and each location are shown in Table 1. The importance of the individual pathways is also exhibited in Figures 1-3 and Table 1. Production rates via

TABLE 1. Instantaneous Peak and 24-h Average Individual and Total HOx Production Rates (million molecules cm-3 s-1) for Each Emissions Scenario for the Three Locations of Interesta for given year CELA

O(1D)

FIGURE 1. Simulated HOx production rates (million molecules cm-3 s-1) versus time on September 9, 1993 in CELA. O(1D)H2O is represented by red, HONOhν is represented by white, H2O2(hν) is represented by purple, HCHOhν is represented by yellow, CARBhν is represented by dark green, ALDNO3 is represented by light green, ALKNO3 is represented by light blue, ALKO3 is represented by dark blue, and ALKO is represented by orange.

peak H2O av O(1D)H2O peak HONOhν av HONOhν peak H2O2(hν) av H2O2(hν) peak HCHOhν av HCHOhν peak CARBhν av CARBhν peak ALDNO3 av ALDNO3 peak ALKNO3 av ALKNO3 peak ALKO3 av ALKO3 peak ALKO av ALKO peak total av total a

FIGURE 2. Simulated HOx production rates (million molecules cm-3 s-1) versus time on September 9, 1993 in AZUS. See Figure 1 for color scheme.

FIGURE 3. Simulated HOx production rates (million molecules cm-3 s-1) versus time on September 9, 1993 in RIVR. See Figure 1 for color scheme. the O(1D)H2O route follow the same spatial and temporal trends (highest downwind and during midday) as the total rates. Conversely, the peak rates associated with HOx formation via the HONOhν route are greatest early in the morning. HONO peaks overnight because of increased relative humidity (reaction between NO2 and H2O to form HONO) at night. A small contribution to HONO formation in the early morning results from the gas-phase reaction of NO and OH before significant amounts of NO are converted to NO2. This contribution is greatest near large-emission sources of NO. Other recent studies have indicated the importance of HONO as a HOx source (21-23). HOx formation

AZUS

RIVR

1993

2010

1993

2010

1993

2010

23.3 4.9 5.3 1.8 0.5 0.1 13.6 3.3 19.1 4.4 1.9 0.5 0.3 0.1 3.3 1.0 0.1 0.0 63.1 16.0

22.7 4.9 2.7 0.8 0.6 0.1 4.6 1.3 7.5 2.1 1.8 0.4 0.5 0.1 0.5 0.2 0.0 0.0 37.5 10.1

33.7 7.1 5.4 1.5 0.3 0.1 14.1 4.4 23.6 7.2 18.5 3.3 2.9 0.6 3.0 1.4 0.1 0.0 78.4 25.5

30.8 6.5 2.1 0.6 0.3 0.1 5.6 1.6 9.9 2.9 8.0 1.4 1.0 0.3 0.8 0.4 0.0 0.0 48.8 13.7

43.4 9.1 7.7 1.2 0.5 0.1 13.5 4.0 24.3 7.2 36.1 5.5 1.8 0.5 2.1 0.9 0.0 0.0 85.6 28.5

32.8 7.2 3.0 0.7 0.5 0.1 5.6 1.7 11.0 3.2 13.6 2.2 0.8 0.2 0.6 0.3 0.0 0.0 51.8 15.6

Values are vertically averaged.

rates via the HCHOhν and CARBhν routes show higher peak values in AZUS and RIVR relative to CELA and are highest during midday. The source from HCHO is slightly higher in AZUS than in RIVR. Such behavior illustrates the nonlinearity associated with formation, emission, and transport of those pollutants such as HCHO that are both primary and secondary. The H2O2(hν) and ALKO routes are insignificant sources of HOx in the SoCAB. The remaining routes of HOx formation exhibit relatively unique behavior compared to the other routes. The ALKO3 route of HOx formation shows the same spatial trend as the HCHOhν route in that peak values are observed in AZUS. However, secondary maxima are also observed in the ALKO3 route late in the day when photolysis decreases yet high concentrations of both alkenes and O3 persist. The ALKNO3 and ALDNO3 routes are relatively insignificant in CELA. In AZUS and RIVR, these routes exhibit peaks late in the afternoon or in the early evening, with the peak rate being greatest for both of these routes in RIVR. These routes are most important in RIVR because they depend on O3 formation to lead to NO3. The rates of formation of HOx from these routes decrease overnight as NO3 and the organics are consumed. In CELA, the maximum peak HOx production is seen for the O(1D)H2O route (23.3 million molecules cm-3 s-1), which also has the highest 24-h average production rate (4.9 million molecules cm-3 s-1). In AZUS, the overall maximum peak is also seen for the O(1D)H2O route (33.7 million molecules cm-3 s-1) despite the CARBhν route having the highest 24-h average rate (7.2 million molecules cm-3 s-1). In RIVR, the O(1D)H2O route shows both the maximum peak (43.4 million molecules cm-3 s-1) and 24-h average production rate (9.1 million molecules cm-3 s-1). Figures 1-3 also indicate the spatial and temporal variation in the contribution of various HOx formation routes to the total HOx formation rate. In CELA, between 0100 and 0600 h and 2000 and 2400 h, the ALDNO3 pathway dominates, although the total formation rates at these times in CELA are very small. During these times, this pathway contributes between 73 and 89% of the total HOx formation rate. In the early daylight hours (0700 and 0800 h), HONOhν is the largest VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

755

source of HOx, constituting between 49 and 53% of the total formation rate. Between 0900 and 1100 h and at 1700 h, CARBhν is the largest source of HOx in CELA with a contribution to total HOx formation of approximately 28%. During the most photochemically active times of the day (1200-1600 h), the O(1D)H2O route is the largest contributor to HOx formation (31-40%). The ALKO3 pathway is the largest individual source (32-38%) of HOx at 1800 and 1900 h. ALDNO3 is the largest source (72-82%) of HOx radical production in AZUS between 0100 and 0600 h and between 1900 and 2400 h. As in CELA, the HONOhν route is dominant in early morning (0700 and 0800 h, 42-45%). CARBhν is again the largest contributor to overall HOx formation during some peak photochemical times (0900-1100 and 1600-1800 h, 30-42%). Between 1200 and 1500 h, the O(1D)H2O route is the greatest, constituting 38-43% of the total HOx formation. These increases relative to CELA indicate the increasing importance of secondary photochemistry with distance downwind. The importance of secondary chemistry is even more emphasized in RIVR, the location of interest furthest downwind, as the importance of NO3 is greatest in this location. Between 0100 and 0600 h, and 1800 and 2400 h, the ALDNO3 route of HOx formation is most important (contributing 2686%). In the early morning (0700 and 0800 h), HONOhν (5263%) is the largest source of HOx, while, between 0900 and 1100 h, CARBhν is of paramount importance (36-44%). As in other locations, the O(1D)H2O route is the largest source during the midday hours (1200-1700 h, 36-52%). The best data to which these simulations can be compared are presented by Paulson and Orlando (7), who used VOC, NOx (NO and NO2), HONO, and O3 concentrations averaged over multiple locations and dates to estimate surface-level mixing ratios needed to calculate HOx production rates in the Los Angeles atmosphere, as in eqs 5a,b,d-f,h. The routes of H2O2(hν) (eq 5c), ALDNO3 (eq 5g), and ALKO (eq 5i) are not considered. Because of the relative insignificance of the H2O2(hν) and ALKO routes, only the ignorance of the ALDNO3 route would be expected to cause any significant differences between the two studies. This difference is important for downwind locations at night, as shown in Figures 2 and 3, but will not affect peak HOx formation rates during the midday. While the temporal behavior of simulated HOx production is similar, lower total HOx production rates (just under 60 million molecules cm-3 s-1) are predicted by Paulson and Orlando (7) compared to the results presented here (range of maximum rates of 63.1-85.6 million molecules cm-3 s-1), despite expecting higher surface-level rates. It must be stated that a direct, quantitative comparison of the results developed here and those of Paulson and Orlando (7) is not possible because of the differences in the methodologies used. The first such difference is that the previous work is based on spatially and temporally averaged ambient concentrations at the surface while the current study is performed using a three-dimensional chemical transport model that incorporates emissions, transport, deposition, chemical reaction, vertical resolution, and other phenomena. While the former method is based on actual ambient observations, the latter approach allows for the ability to reconcile spatial differences within the same airshed and to evaluate the response of HOx formation to changes in emissions. Numerous additional assumptions in the methodologies of both Paulson and Orlando (7) and the present work make direct, quantitative comparison of these two studies impossible. Qualitative comparison of formation rates via individual pathways, however, does lend insight into the differences between the two studies. The O(1D)H2O formation route in Paulson and Orlando (7) constitutes a significantly smaller fraction of total HOx production than in each of the three 756

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 38, NO. 3, 2004

locations simulated here. Such differences are expected because the previous work used a peak O3 mixing ratio of 112 ppb, while simulated peak O3 mixing ratios for the 1993 episode are considerably higher (range of 120-352 ppb in the locations of interest). A similar but much smaller difference is observed for HONOhν. Larger contributions to HOx formation from CARBhν are also observed in the present study in all locations compared to the contribution presented by Paulson and Orlando (7), most likely as a result of the difficulty in identifying all carbonyls present in the atmosphere, particularly those that are part of polyfunctional secondary organic oxidation products. The CIT model is able to predict the mixing ratios of such compounds. The peak HOx formation rate from the HCHOhν route presented by Paulson and Orlando is very similar to the range of peak values in the current modeling work; Paulson and Orlando (7) present peak rates of HOx formation from the ALKNO3 route that are slightly larger than those presented in the current work. Last, Paulson and Orlando (7) indicate significantly higher contributions from the ALKO3 route, despite the higher O3 mixing ratios used in the present study. Higher surface-level concentrations of alkenes used by Paulson and Orlando (7) compared to those predicted in the current modeling would explain the larger contributions from the alkene oxidation routes (NO3 and O3) presented by Paulson and Orlando (7) despite higher mixing ratios of O3 being used in the current work. Sensitivity Study. Because the biggest qualitative differences between the results of Paulson and Orlando (7) and these results are in the ALKO3 route, a sensitivity study is used to identify the influence on modeled HOx formation of reaction rate constants and OH stoichiometric yields in the ALK-O3 reactions. Clearly, this route is most strongly influenced by the mixing ratios of the alkenes, which are most likely higher in the work of Paulson and Orlando (7), as discussed previously. In CACM, lumped parent VOC groups are assigned to represent multiple species. For example, short-chain alkenes of carbon numbers 3-6 are treated as 1-pentene; published kinetic rate constants of 1-pentene (5) and HOx radical yields as predicted by the method of Jenkin et al. (2) are used in the chemical mechanism. However, internal alkenes tend to have higher stoichiometric yields of OH and rate constants for reaction with O3 when compared to 1-alkenes (7). Because the mechanism used in the CIT is a lumped mechanism, the rate constants for the reaction between alkenes and O3 in the mechanism are possibly too small relative to the weighted average O3 reaction rate constants representative of the mixture of alkenes in the SoCAB. However, when model runs were performed in which these rate constants were increased by up to a factor of 10, inappropriate changes in the peakpredicted O3 mixing ratios in the locations of interest were observed. In an effort to investigate the influence on HOx production rates of OH yield, this stoichiometric coefficient was doubled for nonbiogenic alkenes (ethene, short-chain alkenes with three to six carbons, and long-chain alkenes with greater than six carbon atoms). The OH yields were increased because those used by Paulson and Orlando (7) are higher than that obtained by the method described by Jenkin et al. (2). Biogenic alkenes were not considered in this sensitivity study due to their relative unimportance in the SoCAB. The doubling of OH yields did not necessarily lead to a doubling of the ALKO3 HOx production rate since other stoichiometric yields in the gas-phase chemical mechanism also needed to be adapted in order to maintain an atom balance. For example, when 1-pentene reacts with O3 and the ozonide decomposes, the products include HCHO, butaldehyde, and two primary excited Criegee biradicals, one containing one carbon atom and one containing four carbon atoms (5). In the current

model, reaction probabilities for Criegee biradicals as described by Jenkin et al. (2) are used; these include stabilization, which in the case of primary biradicals leads mostly to carboxylic acid formation, decomposition leading to OH formation, and decomposition not leading to OH formation. The third pathway typically leads to carbon dioxide as well as alkyl radicals and hydrogen atoms that form HOx. In the sensitivity analysis, when the reaction probability of OH formation is doubled, the difference is split between the stabilization and non-OH-forming pathways because the sum of the probabilities must equal unity. When the probability of the decomposition pathway that does not lead to OH is decreased, so is the HOx yield in that pathway. Therefore, even though the OH yield is doubled, the overall HOx yield may not be. The results of these changes are a factor of 2 (0.12 to 0.24), 1.68 (0.92 to 1.55), and 1.68 (0.92 to 1.55) increase in HOx production rates from O3 reaction with ethene, shortchain alkenes, and long-chain alkenes, respectively. These represent increases in three of the χi values in eq 5h. These changes were allowed to propagate through the chemistry to influence other photochemical processes. For example, the change in OH yield was allowed to influence OH concentration and the consumption of other VOCs via OH oxidation. Peak O3 in the increased yield case is still underpredicted in AZUS (by 3.1%) and overpredicted in CELA and RIVR (by 13.5 and 40.0%, respectively), performance slightly less favorable on the whole than the base case. The simulated vertically averaged peak production rates of HOx via the ALKO3 route for the base case and the case in which OH yields are increased by a factor of 2 are compared to the peak value (between 25 and 30 million molecules cm-3 s-1) presented by Paulson and Orlando (7). The base case range of peak values is 2.1-3.3 million molecules cm-3 s-1. The range for the increased yield case is 3.3-7.7 million molecules cm-3 s-1. The increased yield values for the surface for this route (up to 11 million molecules cm-3 s-1) are still smaller than those predicted by Paulson and Orlando (7). These results again indicate that significantly higher alkene concentrations must have been used in the study of Paulson and Orlando (7). Emissions Controls. It is of interest to investigate how future emissions scenarios alter peak rates of HOx formation. For this purpose, the emissions scenario for 2010 described in part I (19) is used. Table 1 includes the vertically averaged individual and total peak and 24-h average HOx production rates in each of the three locations for the 2010 scenario. In all cases except H2O2(hν) and ALKNO3 in CELA, simulated HOx production rates in 2010 are smaller than or equal to those predicted for 1993. The slight increase in the H2O2(hν) route is attributed to higher H2O2 levels that result from more efficient HOx-HOx reactions as NOx levels decrease. The increase in ALKNO3 is linked to the slight increase in O3 discussed in part I of this series of papers. With the exception of the ALKO and H2O2(hν) routes, all pathways investigated are significant sources to HOx production in the SoCAB under all of the scenarios presented here. However, all pathways investigated depend on NOx

and VOC emissions, indicating that control strategies that reduce pollutant levels may also alter the route by which some secondary pollutants, particularly HOx, form.

Literature Cited (1) Seinfeld, J. H.; Pandis, S. N. Atmospheric Chemistry and Physics from Air Pollution to Climate Change; Wiley: New York, 1998. (2) Jenkin, M. E.; Saunders, S. M.; Pilling, M. J. Atmos. Environ. 1997, 31, 81. (3) Atkinson, R. J. Phys. Chem Ref. Data, Monogr. 2 1994. (4) Carter, W. P. L. SAPRC-99, available via ftp (http:// helium.ucr.edu/∼carter/), 1999. (5) Atkinson, R. J. Phys. Chem. Ref. Data 1997, 26, 215. (6) Ariya, P. A.; Sander, R.; Crutzen, P. J. J. Geophys. Res., [Atmos.] 2000, 105, 17, 721. (7) Paulson, S. E.; Orlando, J. J. Geophys. Res. Lett. 1996, 23, 3727. (8) Chen, G.; David, D.; Crawford, J.; Heikes, B.; O’Sullivan, D.; Lee, M.; Eisele, F.; Mauldin, L.; Tanner, D.; Collins, J.; Barrick, J.; Anderson, B.; Blake, D.; Bradshaw, J.; Sandholm, S.; Carroll, M.; Albercook, G.; Clarke, A. J. Atmos. Chem. 2001, 38, 317. (9) Comes, F. J.; Forberich, O.; Walter, J. J. Atmos. Sci. 1997, 54, 1886. (10) Mauldin, R. L.; Frost, G. J.; Chen, G.; Tanner, D. J.; Prevot, A. S. H.; Davis, D. D.; Eisele, F. L. J. Geophys. Res., [Atmos.] 1998, 103, 16713. (11) Kanaya, Y.; Sadanaga, Y.; Matsumoto, J.; Sharma, U. K.; Hirokawa, J.; Kajii, Y.; Akimoto, H. J. Geophys. Res., [Atmos.] 2000, 105, 24205. (12) Kanaya, Y.; Matsumoto, J.; Kato, S.; Akimoto, H. J. Geophys. Res., [Atmos.] 2001, 106, 24209. (13) Carslaw, N.; Creasey, D. J.; Heard, D. E.; Lewis, A. C.; McQuaid, J. B.; Pilling, M. J.; Monks, P. S.; Bandy, B. J.; Penkett, S. A. J. Geophys. Res., [Atmos.] 1999, 104, 30241. (14) Poppe, D.; Zimmermann, J.; Bauer, R.; Brauers, T.; Bruning, D.; Callies, J.; Dorn, H. P.; Hofzumahaus, A.; Johnen, F. J.; Khedim, A.; Koch, H.; Koppmann, R.; London, H.; Muller, K. P.; Neuroth, R.; Plassdulmer, C.; Platt, U.; Rohrer, F.; Roth, E. P.; Rudolph, J.; Schmidt, U.; Wallasch, M.; Ehhalt, D. H. J. Geophys. Res., [Atmos.] 1994, 99, 16633. (15) McKeen, S. A.; Mount, G.; Eisele, F.; Williams, E.; Harder, J.; Goldan, P.; Kuster, W.; Liu, S. C.; Baumann, K.; Tanner, D.; Fried, A.; Sewell, S.; Cantrell, C.; Shetter, R. J. Geophys. Res., [Atmos.] 1997, 102, 6467. (16) George, L. A.; Hard, T. M.; O’Brien, R. J. J. Geophys. Res., [Atmos.] 1999, 104, 11643. (17) Grenfell, J. L.; Savage, N. H.; Harrison, R. M.; Penkett, S. A.; Forberich, O.; Comes, F. J.; Clemitshaw, K. C.; Burgess, R. A.; Cardenas, L. M.; Davison, B.; McFayden, G. G. J. Atmos. Chem. 1999, 33, 183. (18) Lu, Y.; Khalil, M. A. K. Tellus, Ser. B 1992, 44, 106. (19) Griffin, R. J.; Revelle, M. K.; Dabdub, D. Environ. Sci. Technol., previous paper in this issue. (20) Griffin, R. J.; Dabdub, D.; Seinfeld, J. H. J. Geophys. Res., [Atmos.] 2002, 107, 4332 (doi: 10.1029/2001JD000541). (21) Calvert, J. G.; Yarwood, G.; Dunker, A. M. Res. Chem. Intermed. 1994, 20, 463. (22) Alicke, B.; Platt, U.; Stutz, J. J. Geophys. Res., [Atmos.] 2002, 107, 8196 (doi: 10.1029/2000JD000075). (23) Alicke, B.; Geyer, A.; Hofzumahaus, A.; Holland, F.; Konrad, S.; Patz, H. W.; Schafer, J.; Stutz, J.; Volz-Thomas, A.; Platt, U. J. Geophys. Res., [Atoms.] 2003, 108, 8247 (doi: 10.1029/ 2001JD000579).

Received for review February 12, 2003. Revised manuscript received October 13, 2003. Accepted October 30, 2003. ES034129V

VOL. 38, NO. 3, 2004 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

757