CURRENT RESEARCH Interim Evaluation of Strategies for Meeting Ambient Air Quality Standard for Photochemical Oxidant Steven D. Reynolds Systems Applications, Inc., 950 Northgate Dr., San Rafael, Calif. 94903 John H. Seinfeld" Department of Chemical Engineering, California Institute of Technology, Pasadena, Calif. 91 125
Two strategies proposed for meeting ambient air quality standards for photochemical oxidant in the South Coast Air Basin of California are evaluated: (1) the 1977 EPA Transportation Control Plan and (2) the current federal light-duty vehicle exhaust emission standards. Currently available means of relating emissions to air quality for photochemical smog are employed in the evaluation. The methods used include modified rollback based on ambient data and laboratory data, the statistical model of Trijonis, and the physicochemical model of Reynolds et al. The Appendix contains a detailed emission inventory projection for Los Angeles in 1977. The 1970 amendments to the Clean Air Act mandated ambient air quality standards for air pollutants that represent long-term goals for air quality in the United States. Perhaps the most controversial aspect of the 1970 amendments is the specified light-duty motor vehicle exhaust emission standards for CO, hydrocarbons (HC), and nitrogen oxides (NOx).As of June 22, 1974, the emission standards have been moved back two years from their original dates. The Energy Supply and Environmental Coordination Act of 1974 specifies the standards as: 1975 Federal 1975 California 1976 Federal 1976 California 1977 Nationwide 1978 Nationwide
HC, g/mi
NO,, g / m i
1.5 0.9 1.5 0.9 0.41
3.1
0.41
CO, g/mi
2.0 . _
15 9
3.1 2.0
15 9
2.0
3.4
0.4
3.4
dards. As noted, we will focus our attention on the photochemical oxidant standard (and, by necessity, also the HC and NO2 standards). Air Quality Models Implicit in the provisions of the Clean Air Act is the supposition that air quality improvements related to emission reductions can be accurately predicted and quantified. In fact, this capability is necessary for rational air quality implementation to be formulated. The methodology of relating pollutant concentrations to emission sources is called an air quality model. The essential elements of an air quality model are source emissions data, meteorological inputs characterizing transport and diffusion processes, and information concerning chemical transformations and removal processes. The type of air quality model, as well as its spatial and temporal characteristics, will be dictated by the purpose for which the model is being employed. Ordinarily, a model is used to determine if airborne pollutant concentrations corresponding to prescribed emission conditions will meet ambient air quality standards. Linear Rollback. For a pollutant which is essentially nonreactive (such as CO) or one which decays according to a first-order chemical reaction (such as SO2 in some instances), the advective-diffusion equation describing the dynamics of the concentration field is linear. Because the basic governing equation is linear, if source emission levels are changed uniformly, with no change in the relative spatial or temporal distribution of emissions, then the resulting change in airborne concentrations (above back-
In 1973 the Environmental Protection Agency (EPA) ( I , 4 ) announced transportation control plans for 23 major
urban areas as a supplementary means of moving more quickly toward satisfying ambient air quality standards. The critical question relating to both new motor vehicle exhaust emission standards and additional transportation controls is: What will air quality levels actually be if these standards and controls are implemented? This paper is addressed to an assessment of the possibility of meeting the national ambient air quality standard for photochemical oxidant. Since the South Coast Air Basin (shown in part in Figure I ) is anticipated to be the most difficult region in which to achieve the photochemical oxidant, HC, and NO2 standards, we confine our attention to the South Coast Air Basin. In particular, using currently available air quality models, we examine the estimated impact on air quality of both the 1977 EPA Transportation Control Plan and the new motor vehicle exhaust emission stan-
Figure Basin
1. Air
pollutant monitoring stations in the South
Coast Air
The 50 X 50 mi area shown I S that considered in the dynamic modeling studies of Reynolds et al (5-7)
Volume 9, Number 5,May 1975 433
ground levels) will be directly proportional to the magnitude of source emission changes. In such a case, it is possible to express the fractional reduction R in all sources from a base year value necessary to meet an air quality standard, D, in a future year as
where P represents a measure of air quality for the base year, B is the background concentration of the pollutant, and g is the growth factor for emissions from the base year to the future year. The form of Equation 1 commonly used, and that which we shall employ, is
R
(gP - D)/(gP- B )
For consistency, the values of P, D, and B in Equation 1 must have the same averaging time. The averaging times for the values of D are stipulated by the air quality standards, and thus the averaging times associated with P and B are determined by those of D. It is important to note that as long as the pollutant is nonreactive or linearly decaying, and as long as emission level changes do not alter the relative spatial or temporal distribution of emissions, Equation 1 is in exact relationship between emission level changes and air quality changes (provided also that B remains constant). Equation l will apply at any point in the airshed as long as P and D refer to values at that point. Although meteorology is not explicitly accounted for in Equation 1, meteorology is reflected implicitly in the measured value of P. The linear rollback fo‘rmula of Equation 1 has been widely used to estimate reductions required to meet ambient air quality standards for various pollutants (8-10). It has been most widely used for CO control. The choice of the base year (present air quality) and the growth factor may have a substantial influence on the calculated emission reduction in some future year. Thus, even if the assumption that the spatial distribution of emissions will not change appreciably is a reasonable one, values of R calculated from Equation 1 may differ owing to different choices of the parameters P and g. In the case of CO, if changes in spatial emission patterns are expected to occur, the rollback model should be replaced by one which accounts explicitly for meteorological conditions ( 1 1 ) . Linear rollback, on occasion, has been applied to estimate reductions in HC emissions to meet the photochemical oxidant standard by assuming oxidant levels are proportional to HC emissions. This is at best a crude approximation since it has been established that oxidant concentrations are nonlinearly dependent on HC and NO, concentrations. Modified Rollback for Photochemical Oxidant. In recognition of the inadequacy of the linear rollback method for representing the link between precursor emissions and resulting oxidant concentrations, there has been considerable effort devoted to finding means by which the observed nonlinear oxidant-precursor relationship might be characterized by the basic rollback approach. The method arrived at might be termed modified rollback. In this approach, peak oxidant levels are first related to initial HC (and possibly NO,) levels as determined from either ambient monitoring data or laboratory smog chamber data. It is then assumed that the initial HC and NO, concentrations are linearly proportional to HC and NO, emissions, and that these concentrations are linearly “rolled back” by reductions in HC and NO, emissions. Oxidant-Precursor Relationships Based on Ambient Monitoring Data. An approach adopted by EPA in 1970 is to equate the highest observed oxidant value on a par434
Environmental Science & Technology
ticular day with the 6-9 a.m. average hydrocarbon concentration (expressed as nonmethane hydrocarbon) measured on that day. The data points for all available days are then plotted in the plane of maximum 1-hr average oxidant and 6-9 a.m. average HC. A curve is then drawn through each of the largest oxidant values at each HC value. Such a plot, which might be called an upper limit curve, has been prepared by Schuck and Papetti (12) for maximum 1-hr average oxidant observed anywhere in the South Coast Air Basin as a function of 6-9 a.m. HC concentration averaged over 8 stations (Figure 2 ) . Although NO, levels are not explicitly accounted for, the data through which the upper limit curve was drawn, reflected a variety of NO,/HC ratios. For a particular year it might be expected that the NO,/HC ratio would be fairly uniform from day to day at a particular location. The explanation as to why the NO,/HC ratios observed on different days during 1971 should vary appreciably is not clearly evident. Differing wind directions and temperatures, with the effect of the latter on HC volatility, are two possible causes. A drawback of Figure 2 is that post-9 a.m. emissions are unaccounted for. Oxidant-Precursor Relationships Based on Laboratory Data. A shortcoming of the upper limit analysis as in Figure 2 is the lack of explicit inclusion of NO, levels. It would be desirable to extend Figure 2 in such a way that curves of constant peak oxidant were plotted in the plane of initial HC and NO, levels. Unfortunately, a correlation of this type does not appear to be available for the South Coast Air Basin. Thus, in order to include both HC and NO, levels, it has been necessary ’to rely on the results of laboratory experiments with actual auto exhaust, with realistic irradiation times, and at realistic concentrations, to generate two-dimensional maximum oxidant surfaces in the plane of initial HC and NO, concentrations, Perhaps the most relevant set of laboratory experiments for determining oxidant-precursor relationships in ambient air is that reported by Dimitriades (13, 14) and summarized in Figure 3. The sloping lines of Figure 3 correspond to the National Air Quality Standard for oxidant (0.08 ppm 1-hr average). Values of HC and NO, to the left of line a b and below bc yield less than 0.08 ppm oxidant after 6 hr of irradiation equivalent to Los Angeles sunlight. Dimitriades included line df to represent an estimate of the 3-hr average NO, concentration equivalent to the federal air quality standard for NO2 (0.05 ppm, annu-
6-90m Averoge Non Methone Hydrocorbon Concentrotlon IppmCI
Figure 2. Maximum 1-hr average oxidant concentrations observed at 12 stations in the South Coast Air Basin as a function of the 6:OO-9:00 a.m. nonmethane hydrocarbon concentrations observed at 8 stations during 1971 (72)
a1 average). In the cross-hatched regions, neither the oxidant nor NOz standards are violated after 6 hr of irradiation. The use of laboratory results, such as those in Figure 3, for determining atmospheric oxidant-precursor relationships has been criticized. One factor is that the duration of time over which the smog chamber experiments were carried out is usually a t most 6 hr, although a t noonday intensities. Less reactive hydrocarbons can be oxidant precursors after long periods of time when the more reactive hydrocarbons have disappeared. Perhaps a more important drawback of most smog chamber data for atmospheric interpretations is the lack of inclusion of continuous fresh emissions. Oxidant-Precursor Relationships Based on Statistical Analyses. For given HC and NO, emission levels, Trijonis (15) determined the joint distribution of morning HC and NO, concentrations (7:30-9:30 averages) from five years of APCD monitoring data (1966-70). Then, on the basis of oxidant and NO2 data over the same five-year period, he determined the probability that the l - h r California oxidant and NO2 standards of 0.10 and 0.25 ppm, respectively, would be violated as a function of morning concentrations. For oxidant, an average was taken of maximum 1-hr values between 11:OO a.m. and 1:OO p.m. in downtown Los Angeles, Burbank, and Pasadena, weighted according to wind speed and direction, so that the maximum oxidant would correspond as closely as possible to that in the air mass that had been over the downtown area in the morning. Separate analyses were performed for winter and summer. The overall result of Trijonis’ analysis is Figure 4, in which the expected number of days that the California oxidant and NO2 standards are violated is shown as a function of total daily HC and NO, emissions in Los Angeles County. There are several shortcomings inherent in using the statistical model of Trijonis for planning purposes. First, only the downtown Los Angeles, Burbank, and Pasadena stations are considered. The highest oxidant levels occur significantly east of downtown Los Angeles. Thus, satisfaction of the oxidant standard at the downtown station will not ensure that oxidant violations will not occur downwind of central Los Angeles. Second, the emissions data on which the model is based (that for 1966-70) exhibited only modest changes in total HC and NO, emissions. For future prediction purposes, one is interested in NO, and HC levels well below those of 1966-70. Use of the correlation under these much lower emission conditions is highly uncertain. Physicochemical Models. The problem of depicting the emission level-air quality relationship for photochemical smog could be resolved with the use of a mathematical model of proved reliability, based on an accurate description of atmospheric transport and meteorological and chemical processes. Because of the difficulties in developing and incorporating nonlinear chemical mechanisms into advection-diffusion models, the development of physicochemical models for reactive pollutants has lagged behind that of models for inert pollutants. Nevertheless, a comprehensive photochemical-diffusion model has been tested and validated in a preliminary fashion for Los Angeles (5-7, 1 6 ) . The region considered by Reynolds et al. ( 5 )in their studies is shown in Figure 1. It is important to note that current physicochemical models predict the dynamics of pollutant concentrations over the course of a particular day or perhaps several days. Air quality standards have been expressed, on the other hand, in terms of a frequency of yearly violations. Thus, there is no direct way to compare the predictions of
a dynamic model to current air quality standards. One possible approach is to select the day (i.e., meteorological conditions) on which the highest oxidant level was reached in a given year and to use the same meteorology to test the effect of emission level changes. Organization of the S t u d y
The main objective of this study is to compare the predictions of available methods of relating emissions to air quality in photochemical smog when applied to two problems: (1) assessment of the effectiveness of the 1977 EPA Transportation Control Plan in meeting ambient air quality standards for oxidant, CO, NOZ, and HC in the South Coast Air Basin, and (2) assessment of the effectiveness of the current motor vehicle exhaust emission standards in ultimately meeting the same objective. It is recognized that the specific EPA control plan considered here may soon be obsolete; however, the types of measures in the plan may well be included in future strategies. Thus, the analysis to be presented. here will hopefully provide a framework for the assessment of future air pollution control plans. The first step in the evaluation of control strategies is the compilation of a complete contaminant emissions inventory for the region over the time period of interest. The inventory should include all major sources, their locations, and, if a dynamic model is to be used, the temporal distribution of their emissions. In compiling an inventory for future years, three elements must be considered carefully:
Slope
0605-
E
(Current atmospher c levels) g
8 04-
0 0
2:o 3.0 4.0 Total Hydrocorbon, ppmC 0.94 1.88 2.82 376 Non-Methane Hydrocorbon, ppmC I .0
-5.0 4.70
Combinations of initial total reactive hydrocarbon and concentrations corresponding to oxidant and NO2 yields equal to national ambient air quality standards (73) Figure 3.
NO,
LL
o
0.2 0.4 0.6 , z 02 04 06 08 IO 1.2 Fraction ooff I969 R H C Emissions i n Las Angeles C o u n t y
~
The expected number of days per year that 1-hr average oxidant and NO2 levels in downtown Los Angeles exceed the California ambient air quality standards of 0.10 and 0.25 p p m , respectively, as a function of total daily emissions of R H C a n d NO, in Los Angeles County ( 7 5 ) Figure 4.
Volume 9, Number 5, May 1975
435
(1) the expected rates of growth of various sources and how these rates vary with location in the airshed, (2) the expected controls to be implemented between the present time and that under consideration, and (3) the expected performance of these controls. To evaluate the 1977 EPA transportation control plan, two complete emission inventories are developed for the Los Angeles airshed, including a nominal (or baseline) inventory for 1977, and an inventory obtained by modifying the baseline emissions through implementation of the EPA control strategies. Emissions in the baseline inventory are intended to be representative of those expected in the Los Angeles airshed (considered here as the 50 x 50 mi region shown in Figure 1) in 1977 if no additional control measures other than scheduled motor vehicle exhaust emission standards on 1975-77 vehicles are implemented. The two 1977 inventories are developed in detail in the Appendix, where we present a general methodology for the compilation and updating of an urban emissions inventory with particular attention to motor vehicle emissions. Once the inventories have been developed, each of the air quality modeling approaches can be used to estimate the degree of air quality improvement from implementation of the EPA transportation control plan. Finally, using the two modified rollback approaches, we assess the ultimate effectiveness (in 1988) of the current Federal light duty vehicle exhaust emission standards. E P A Transportation Control P l a n
The EPA control plan for Los Angeles consists primarily of several actions to reduce vehicular emissions. The plan is composed of the following elements ( 4 ) : 1. Limitation of annual gasoline sales in years subsequent to 1973 to that amount sold between July 1, 1972, and June 30, 1973 2 , Mandatory periodic vehicle inspection and maintenance of all light-duty vehicles 3. Establishment of exclusive bus and carpool lanes on freeways and major streets during the morning and evening peak traffic hours 4. Restricted registration of all motorcycles to 1973 levels beginning in 1974, and restriction of the operation of two-stroke motorcycles during daylight hours in the months from May through October, beginning May 1, 1974 5. Installation of vacuum spark advance disconnect (VSAD) retrofit devices on pre-1971 light-duty motor vehicles, and installation of oxidizing catalyst retrofit devices on 1966-74 light-duty motor vehicles. The plan also contained originally provisions for a ban on construction of new parking facilities and a 20% reduc-
tion of existing facilities. Congress subsequently banned EPA from regulating parking spaces, so we will not consider this aspect of the plan. While specific control measures for fixed sources were not included in the EPA plan, EPA is considering such measures as substitution of unreactive solvents for reactive solvents and use of vapor recycle systems. Liu (17) has recently compiled a fixed source inventory for the Basin in 1977 assuming implementation of control measures similar to those considered by EPA. We have assumed that these measures are included in the EPA plan outlined above. Although it is not clear precisely how these measures were arrived at, it appears that they represent a group of technologically feasible steps which would serve to reduce total mass emissions of CO, RHC, and NO, by percentages which, by means of a rollback calculation, would reduce pollutant levels to those closely approaching the air quality standards. Two emission inventories are of interest in this study: (1) the 1977 baseline inventory representing emissions of CO, hydrocarbons, and NO, into the Basin assuming no further controls other than those required of pre-1978 model year motor vehicles, and (2) the 1977 inventory assuming implementation of the EPA control strategy. The methodologies employed to estimate motor vehicle, aircraft, and fixed source emissions in both cases are presented in the Appendix. A complete summary of the total estimated daily emissions for 1977 in the modeling region is given in Table I. As an aid in comparing the baseline inventory with that for 1969 and in comparing the inventory resulting from implementation of the EPA control strategy with the 1977 baseline and 1969 inventories, percentage changes are shown in Table I. Note that RHC and URHC refer to reactive and unreactive hydrocarbons, respectively, while URHC includes methane, ethane, propane, benzene, and acetylene, and RHC is comprised of all other hydrocarbons. Evaluation of 1977 EPA Control Plan
We now wish to evaluate the probable air quality that might result in 1977 both with and without implementation of the EPA control plan. Having determined emissions inventories for 1977 both with and without the implementation of the EPA control plan, we can assess by means of modified rollback approaches and physicochemical modeling, the estimated yearly maximum 1-hr average oxidant concentration in the Basin in 1977. The results based on modified rollback are summarized in Table 11. We include predictions based on linear rollback of oxidant by HC reduction for comparison only, since it is not a rec-
Table I. Total Daily Projected Emissions in Los Angeles Modeling Region in 1977, Tons/Day NO,(as NOZ)
RHC
co
URHC
Source
1977W
1977Eb
19778
1977E
Motor vehicles Aircraft Power plants Oil refineries Misc. stat. sources Total emissions Change from 1969,% Change from 1977B,%
406 6 159 88
314 6 159 34
444 26
322 26
52 6
28 6
34
14
37
31
1977 baseline.
436
134
1977E
19778
1977E
2882 79
1831 79
-
-
-
640
115
246
246
-
-
-
793
575
1114
477
341
311
2961
1910
-64
-77
-8
62
19175
-33
-28 1977 EPA control strategy.
Environmental Science i 3 Technology
-39
-
-75 -58
-
-25
-31 -9
-35
ommended procedure for photochemical oxidant. Both results do not predict compliance with the oxidant standard if the strategy is enacted. Next we wish to employ the comprehensive dynamic simulation model of Reynolds et al. (5-7). As noted, in using a dynamic model, the meteorology for a particular day must be chosen as inputs to the calculation. The dynamic simulations to be performed for the 1977 emission levels are to be based on a day with the meteorology of September 29, 1969. Thus, wind speed and direction, mixing depth, turbulent diffusivity, and solar radiation intensity inputs are identical to those used in the model evaluation studies for this day (6). The remaining inputs to be specified are reaction rate constants and initial and boundary concentrations. The only reaction rate constants possibly requiring revision from the previously reported 1969 simulations are those associated with reactive hydrocarbons, particularly if changes occur in the types of hydrocarbons emitted from sources. IJnfortunately, detailed information with regard to the hydrocarbon composition of new and future automobile, aircraft, or fixed source emissions has not been reported. Thus, it will be assumed for this study that the relative amounts of each reactive hydrocarbon species are unchanged from those found in measurements taken on September 29, 1969, and therefore that all rate constants and stoichiometric coefficients for reactions involving RHC are the same as those used in the 1969 evaluation studies (6). The initial and boundary concentrations employed for the 1977 simulations were obtained by modifying those used in the September 29, 1969, validation. Thus, to facilitate the comparison of the 1969 and 1977 model results, the objective is to estimate the initial and boundary conditions that would have resulted on September 29, 1969, had emissions been a t those levels estimated for 1977. Since CO and URHC are essentially inert, it is assumed that initial and boundary concentrations of these two species are reduced from the 1969 values in direct proportion to the reduction in total emissions of CO and URHC in 1977. Lacking better guidelines, it is also assumed that RHC and total NO, (i.e., NO NO2) concentrations appearing in the initial and boundary conditions are reduced from their 1969 values in proportion to the 1977 emission reductions. The key problem is estimating the ratio of the concentrations of NO to NO2 and the ozone concentration in the initial and boundary conditions for the 1977 studies. Since the initial and boundary NO/N02 ratio and ozone concentration prove to have an important effect on predicted concentrations throughout the day, two cases were studied that reflect the limits that these concentrations might be expected to obey. In prior model evaluation studies, ozone initial and boundary concentrations have been calculated from the ozone steady state relationship so that, if this relationship is assumed to hold, specifying the ratio of NO2 to NO fixes the ozone concentration. Case 1. The first method of calculation is based on the assumption that NO and NO2 concentrations at sunrise and at the boundaries of the airshed will be reduced from the 1969 values in direct proportion to the reduction in total NO, emissions in the Basin-i.e.,
+
[NO];, = +NOx[N0169 [NO2177 = @NO,lN02169
in 1969. If a steady state relationship is assumed to exist among NO, NO2, and 0 3 , the ozone concentration may be approximately computed from (in the exact expression there is a small dependence of [O3lSson [RHC]) (4)
-
(5)
where ~ R H Cis the fractional reduction in RHC emissions in 1977 from 1969. Typical NO, NOz, and 0 3 concentrations predicted in Cases 1 and 2 are presented in Table 111. Case 1 predicts a ratio of [NO21 to [NO] four times as large as Case 2, and consequently an ozone concentration four times that of Case 2. Clearly, the initial and boundary concentrations that will occur in 1977 are the result of very complex pro-
Table II. Maximum Observed or Predicted 1-Hr Average Oxidant Concentrations in South Coast Air Basin Year
Upper-limit cu rve (Fig. Z ) , p p m
Statistical model (Fig. 4)a
Linear rollback, ppm
1969 1977 baseline 1977 EPA
0.55b
150'~
0.55b
0.44 0.21
110 10
0.39 0.14
a Daysjyr 1-hr. average oxidant in central Los Angeles exceeds 0.10 gpm. We have assumed t h a t t h e percentage reduction in emissions of H C a n d NO, from Table I for t h e 50 mi. X 50 mi. region of Figure 1 also apply to 10s Angeles County. Observed.
Table Ill. Typical NO, NO,, and O3Concentrations (Ppm) Predicted Along the Palos Verdes Coastline at Noon by Cases 1 and 2
(2) (3)
where the subscripts 69 and 77 refer to the initial or boundary concentrations in 1969 and 1977, and C++NO, is the ratio of the total daily NO, emissions in 1977 to those
-
where 121 and 123 are the rate constants for the two reactions: NO2 + hv NO + 0 and NO + 0 3 NO2 + 0 2 . From Equations 2 and 3, it is seen that the ratio of [NO21 to [NO] does not change from 1969 to 1977 in this case, and consequently that the initial and boundary ozone concentrations will be virtually unchanged from 1969 to 1977. Thus, this method of calculation does not lead to lower ozone concentrations at sunrise or outside the airshed in 1977 in spite of the reductions in NO, and RHC emissions. This method of calculation should provide an upper limit on the estimated initial and boundary ozone concentrations in 1977. Case 2 . The second approach for specifying the NO, NO2, and 0 3 initial and boundary conditions is suggested by the rollback concept, in which ozone concentrations are assumed to be proportional to RHC emissions and total NO, concentration, either initially or a t the boundary, is assumed to be proportional to NO, emissions. The following two expressions may then be written for [N0I77 and [NO2177 (again assuming that ozone obeys the steady state relationship):
Case 1 Case 2
[Noh7
[NO?lii
0.0034 0.0097
0.0207 0.0146
[NOnIiil [NOIri
6.0
1.5
[Oalii
0.1228 0.0304
NOTE: Concentrations in this table were derived for t h e 1977 EPA control strategy. From Table I , @RHC = 0.25 a n d @so, = 0.67. Values for t h e rate constants used were: k l = 0.44 min-1. k3 = 21.8 ppm-1 min-1.
Volume 9,Number 5, May 1975
437
cesses, such as the previous day’s conditions and the meteorology at night. These two simple methods of estimation certainly do not take these factors into account. It is to be stressed that the basic object of these two cases is to illustrate the range of values that can be obtained using “reasonable” assumptions. In this section simulation results will be presented corresponding to use of both Cases 1 and 2. These simulations indicate the level of sensitivity RESEOR
I “
-
PREOlCTEO
CO
\
I
\
5
6
7
8
9
10
11
12
13
1U
TINE ( P S T )
DOWNTOWN LR
I
15
5
7
6
8
9
10
11
12
13
1V
TINE ( P S I )
Figure 5. Predicted hourly averaged CO concentrations at Reseda and downtown Los Angeles. On this figure and those to follow, A , B, and C designate predictions employing the 1969, 1977 baseline, and 1977 E P A control strategy emissions inventories, respectively
of the computed concentrations to the initial and boundary conditions. Combining the various meteorological, emissions, and chemical inputs, simulations were carried out for the 5 a.m.-3 p.m. PST period for each of the two emission inventories discussed previously and for each of the two initial and boundary concentration algorithms. Computergenerated plots of predicted hourly averaged CO concentrations at Reseda and downtown Los Angeles are presented in Figure 5 . On Figure 5 and subsequent plots the letters A , B, and C designate predictions employing the 1969, 1977 baseline, and 1977 EPA strategy emissions inventories, respectively. Predictions of NO, NOz, and O3 concentrations are given in Figures 6a to 7b. In each set, the figures labeled “a” and “b” denote the use of Cases 1 and 2, respectively, for the initiallboundary condition computation. The maximum O3 concentrations predicted for the EPA strategy range between 0.04 and 0.14 ppm, depending on the algorithm used to compute the NO, initial and boundary conditions. Finally, it must be noted that the region considered here extends on its eastern boundary only to Pomona. Since the highest oxidant levels occur often at Riverside and San Bernardino, one cannot draw conclusions about maximum O3 levels without extension of the region to include these areas. A direct comparison of the dynamic model predictions with those of Figure 2 is not appropriate, because Figure 2 is based on the highest observed oxidant at any given hydrocarbon level, and the highest oxidant level during 1969 was not observed on September 29, 1969. In addition, a direct comparison of the dynamic model results with those of Figure 3 is not possible because it is not clear how the initial hydrocarbon and NO, concentrations on the axes of Figure 3 should be computed for the airshed. Maximum 1-hr average ground-level concentrations of CO, NOz, and 0 3 predicted on a day with the meteorology of September 29, 1969, for the 1969 and two 1977 inventories are presented in Table IV together with those predicted on the basis of linear rollback from the 1969 values. Because CO is inert and because the spatial distriBURBRNK
BURBRhiK
-
I
“1
A
PREOICTEO
NO
- PRLOJCTEO --
uo -
PREOICTEO
A
I
I
NO
NO2
A
$” 30-c U
IO-
_ - B_ _ 5
6
7
8
3s
-
11
12
13
1U
BURBRNK
- PREOJCTEO
30
=I
03
25-
”
I
A/ 5
5-
5
6
7
U
9
10
I1
12
13
1U
TINE (PST)
Figure 6a. Predictions of hourly averaged NO, NO2, and O3 concentrations at Burbank. Case 1 initial and boundary conditions 438
10
BURBRNK
t
I k 20 20V I V
9
TINE (PST)
TINE ( P S T )
Environmental Science & Technology
5
6
7
U
9
10
I1
12
13
1‘4
TIME (PST)
Figure 6b. Predictions of h o u r l y averaged NO, NO?, and 0 3 concentrations at Burbank. Case 2 initial and boundary conditions
bution of emissions in the Basin is not expected to change radically from 1969 to 1977, good agreement between the dynamic model and linear rollback predictions can be expected. For NO2 and 03,the maximum concentrations predicted on the basis of linear rollback are generally higher than those predicted by the dynamic model. With the exception of the EPA inventory using the Case 1 IC/BC algorithm, the 0 3 predictions from the dynamic model are in most cases considerably lower than the corresponding rollback predictions. The agreement in that one case is probably coincidental since the maximum O3 concentration of 0.14 ppm occurred adjacent to the boundary near the coast and is a direct result of the boundary condition used.
Table IV. Comparison of Maximum 1-Hr Average Ground-Level CO, NO,, and O3 Concentrations Predicted by Linear Rollback and Dynamic Model. 1977 Baseline
1969 DM
Rb
17
6.4
DM1
1977 EPA strategy
DM2
Rb
DM1
DM2
4.3
4
4
76
76
co
Max. concn, PPm
Reduction from 1969, %
62
6
6
65
65
75
0.36
0.35
0.26 0.15 0.09
5
8
NOn
Max. concn, PPm
0.38 0.35
Reduction from 1969, %
8
60
33
76
0 8
Max. concn, PPm
0.54 0.34 0.19
Reduction
from 1969, % '
38
0.16 0.15 0.14 0.04
65
70
74
72
96
Results under t h e Rb column are based on t h e linear rollback as. sumptions. results under t h e DM1 a n d DM2 columns are based on prediction; from t h e dyna,mic model using t h e first a n d second NO, I C/BC algorithms, respectively. a
t
35
30
To partially explain the relatively low O3 predictions obtained from the airshed model, note in Table I that the percentage reduction in RHC emissions for both 1977 inventories is greater than the corresponding reduction in NO, emissions. When compared to the 1969 results, smaller conversions of NO to NO2 are expected in the 1977 simulations (over comparable reaction times) resulting in the formation of lower O3 concentrations. Furthermore, use of the second IC/BC algorithm rather than the first cause a relative increase in the ratio of NO to NO2 concentrations equal to the initial and boundary conditions which also results in lower predicted O3 concentrations. The observations cited above are illustrated in Figures 6a and 6b. Since the choice of initial and boundary concentrations has an appreciable effect on 0 3 predictions during the first few hours of the simulation, particular attention must be given to the specification of these parameters, especially for future years when a considerable change in source emissions is expected to occur. To minimize the influence of initial conditions, multiple day simulations may be performed. In this case, the results predicted on the second or subsequent days may be used to evaluate the control strategy. Uncertainties introduced through the need to specify pollutant concentrations at points of horizontal inflow to the region may be reduced by including the urban area plus a sufficient portion of the surrounding areas in the region so that background levels may be assumed at the inflow boundaries. We note that there is considerable discrepancy between the predicted oxidant reductions from linear rollback, modified rollback (Figure 2), and the dynamic model. When compared with Figure 2, linear rollback tends to overestimate the percentage reduction in maximum oxidant concentration, but underestimates compared with the dynamic model. Peak ozone levels predicted by the dynamic model may be in error by as much as a factor of two (6). Thus, it is difficult to draw any firm conclusions from the analysis. Maximum hourly ozone levels could
RZUSR 35
- PREDICTED - - PREDICTEO
t
c
RZUSA
4
NO NO2
- PREDICTED
NO
- - PAEOICTEO
NO2
C.
5
6
7
8
9
TIME
6o
t
10
I1
12
13
14
(Psr)
TIME (PST)
RZUSR
RZUSR
-
PRLOlCTEO
- PREDICTED
03
03
5 uo-
M -
5
6
1
0
9
ID
11
12
13
1U
TIME (PST)
Predictions of hourly averaged NO, NO*. and 03 concentrations at Azusa. Case 1 initial and boundary conditions
Figure 7 a .
Predictions of hourly averaged NO, N02, and 03 concentrations at Azusa. Case 2 initial a n d boundary conditions Figure 7b.
Volume 9 , Number 5, May 1975
439
probably be expected to lie in the range of 0.10-0.15 ppm if the EPAstrategy were implemented.
Evaluation of HC and NO, Exhaust Emission Standards One of the most controversial aspects of the Clean Air Act is the motor vehicle exhaust emission standards for HC and NO, necessary .to enable meeting of ambient air quality standards for HC, NOz, and photochemical oxidant. Although the current standards will be enforced in a stepwise manner until 1978, the most important question is whether the 1978 standards of 0.41 g/mi HC and 0.4 g/mi NO, will enable ultimate conformance with the ambient air quality standards for HC (0.24 ppmC, 3-hr average), NO2 (0.05 ppm, annual average), and photochemical oxidant (0.08 ppm, 1-hr average). In this section we apply the current techniques available for making such an assessment to evaluate the federal exhaust emission standards. The available techniques which can be used for this assessment are modified rollback based on ambient monitoring data (the upper limit curve of Figure 2), modified rollback based on laboratory smog chamber data (Figure 3), and the physicochemical model. In principle, the physicochemical model is the preferred approach of those available. The use of such a model does require, however, the compilation of a complete spatio-temporal emission inventory, a task which we have not undertaken for Los Angeles in the target year of 1988. Thus, we shall present only the results based on modified rollback, with the proviso that these results should be reconsidered when appropriate emission inventories are prepared for the target year so that physicochemical models can be employed. While the ambient air quality standards for NO2 and oxidant were established on the basis of health effects, that for HC was set to ensure meeting of the oxidant standard. Both Figures 2 and 3 can be used to estimate the 6-9 a.m. average nonmethane hydrocarbon (NMHC) concentration required to yield oxidant levels less than 0.08 ppm. Figure 2 predicts that a value of the 3-hr average NMHC concentration necessary is about 0.3 ppmC, at presumably the most unfavorable NO,/HC ratio at that level in 1971. Figure 3 takes explicit account of both initial HC and NO,. The necessity of reaching the shaded area d e b of Figure 3 will determine the degree of control of both HC and NO, required. It should be noted that reducing the HC level to point e, the oxidant standard will be met only if NO, is also a t point e. If NO, is below level e, the oxidant level will exceed the standard. Dimitriades (14) has noted that since the NO, corresponding to a certain HC level varies by as much as 6070, the HC must be reduced below the e level to ensure low oxidant levels at all NO, levels. Based on Figure 3, Dimitriades has estimated that the NMHC must be reduced to 0.20-0.25 ppmC to ensure attainment of the oxidant standard regardless of NO, variation. Thus, both ambient data (Figure 2) and laboratory data (Figure 3) appear to confirm the current federal ambient air quality standard of 0.24 ppmC (6-9 a.m. average). We can now use the desired air quality level of 0.24 ppmC for NMHC and the generally accepted background level of 0.1 ppmC to study the effect of various values of P and g on the percent reduction in HC emissions. We take 1970 as the base year and 1988 as the target year. Figure 8 shows the percent reductions and required exhaust emission standards for the four growth factors considered by the Panel on Emission Standards and Atmospheric Chemistry (18) as a function of present air quality P. The 1970 baseline weighted HC auto emissions have been computed as 10.8 g/mi [hot start rate based on Federal Driving Cycle, measured in 1971 (16)]. The highest 6-9 a.m. 440
Environmental Science & Technology
NMHC concentration observed during 1970 in Los Angeles was 4.7 ppmC. The dashed line at P = 4.7 indicates that the required HC emission standard is always below 0.41 g/mi for growth Tactors from 1.0-1.9. Thus, the results indicate that if a growth factor of HC emissions greater than 1.0 can be expected for Los Angeles from 1970-88, the present exhaust emission standard of 0.41 g/mi may not be sufficiently stringent to ensure compliance with the oxidant standard. In using the rollback method, we assume that stationary sources are to be controlled to the same extent as vehicular sources. A similar calculation can be carried out for the NO, exhaust standard. The national ambient air quality standard for NO2 is 0.05 ppm/annual average. If a value for the annual average NO2 concentration is available for the base >ear and region of interest, then that value can be used for P with D = 0.05 ppm. Alternatively, the annual average value of 0.05 ppm can be translated into 1-hr or 3-hr maximum values. There are two basic ways to carry out this translation. First, the concentration data can be assumed to follow a given frequency distribution (generally log normal). Assuming NO2 levels are log-normally distributed, the Panels on Emission Standards and Atmospheric Chemistry (18) estimated the 1-hr equivalent of a 0.05 yearly mean as 0.35 ppm. Second, from ambient data for various cities, the ratios of the maximum 1-hr average NO2 to the average annual NO2 can be computed. Presumably one might then choose the largest ratio as the basis for determining the 1-hr NO2 maximum. Preliminary calculations using this latter approach indicate that Los Angeles is generally the region with the largest ratio and that the 1-hr maximum NO2 is about 0.35 ppm, in
c
I
2 3 4 5 (P) Present Air Quality (pprnC, 3-hr. average)
Figure 8. Fractional reduction in hydrocarbon emissions based on the federal ambient air quality standard of 0.24 ppmC, 3-hr average, as determined by linear rollback. The emission standard is computed based on the 1970 vehicle population estimated emission level of 10.8 g / m i
D = 0.30 ppm (!-hr. average) B = 0.03 ppm
I O 04
la
..-. E
05
06 07 08 09 IO I I 12 'P) P r e s e n t A i r Q u a l i t y (ppm, I-hr o v e r a g e )
9. Fractional reduction in NO, emissions based on the 1-hr equivalent to the federal ambient air quality standard for 'NOa of 0.05 ppm, annual average, as determined by linear rollback. The emission standard is computed based on the 1970 vehicle population estimated emission level of 4.16 g / m i
agreement with that calculated from the assumed statistical distribution (19). Once the desired NO2 value has been determined, that value must be converted into an NO, concentration. For this conversion ambient data are needed. Recent examination of Los Angeles monitoring data indicates that the maximum 1-hr NO2 concentration is generally equivalent to the maximum 1-hr NO, concentration (19). By similar procedures the Panels on Emission Standards and Atmospheric Chemistry (18) have estimated that the lowest 3-hr NO, value equivalent to the NO2 standard is 0.30 ppm. Figure 9 shows the percent reduction for the four growth factors of Figure 8 as a function of present air quality (measured as 1-hr NO, average) for a desired level of 0.30 ppm, a background value of 0.03 ppm, and for a 1970 baseline weighted emissions value of 4.16 g/mi [measured in vehicle surveillance studies in Los Angeles in 1971 (16)). A typical high 1970 value for NO, was 1.0 ppm. With that value, it appears that the current NO, standard may be somewhat stringent based on meeting the NO2 standard. However, it is not possible at present to assess the NO, emission standard that is required, together with the HC emission standard, to ensure that all regions of the South Coast Air Basin will not exceed the oxidant standard.
Summary Two strategies for achieving the ambient air quality standard for photochemical oxidant have been examined:
(1) the 1977 EPA transportation control plan for the Los Angeles airshed, and (2) the federal light-duty motor vehicle exhaust emission standards for HC and NO,. Results of the evaluation of the EPA transportation control plan indicate about a 75% reduction in peak CO, from a 33 to 600/0 reduction in peak NOz, and from a 60 to 70% reduction in oxidant from 1969 levels. A preliminary analysis of the effectiveness of the 1978 federal exhaust emission standards for HC and NO, was carried out. The analysis indicates that the 1977 HC standard of 0.41 g/mi may not be stringent enough for meeting the ambient oxidant standard in the Los Angeles Basin. While it was found that the 1978 NO, standard of 0.4 g/mi may be somewhat stringent for meeting the ambient NO2 standard in Los Angeles, not enough information is available at present on the relationship of NO, emissions to oxidant air quality to warrant a recommendation that the NO, emission standard be loosened. Further work is urgently needed on this aspect.
Acknowledgment Assistance in the preparation of the motor vehicle and fixed source emissions inventories by M. K . Liu is gratefully acknowledged. A portion of this analysis was carried out in conjunction with preparation of the report, “Air Quality and Automobile Emission Control,” Vol. 3, “The Relationship of Emissions to Ambient Air Quality,” National Academy of Sciences, August 31, 1974.
APPENDIX. PROJECTED EMISSIONS FOR LOS ANGELES IN 1977 1977 Baseline Emissions Inuentory-Motor
Vehicles Since motor vehicles will contribute significantly to the photochemical air pollution problem in the Los Angeles Basin for many years, a comprehensive emissions inventory applicable during this period must include exhaust and evaporative emissions from these important sources. The two quantities needed to describe the total exhaust emissions injected into the airshed are (1) the rate a t which each pollutant is emitted from the “average” motor vehicle (g/mi), and (2) the number of miles traveled on both surface streets and freeways in each ground-level grid cell (mi/day). Because driving conditions on surface streets and freeways tend to differ, the rate a t which pollutants are emitted from a vehicle will depend on the type of roadway on which it is operated. Driving on surface streets involves a full range of idle, cruise, acceleration, and deceleration operating modes, and thus emissions are best quantified by measuring the amounts of pollutants emitted while the vehicle is operated on a driving cycle, such as the Federal Driving Cycle (FDC). It is also important to have FDC emission estimates based on both hotand cold-start tests. On freeways, however, vehicles may be subjected to somewhat more stable operating conditions (i.e., the driving speed tends to fluctuate less). Emissions may be quantified in this instance by measuring the rate at which pollutants are emitted at several cruising speeds and from this data, determine an emissions/ driving speed correlation. In this section the discussion will focus on the methodology employed to calculate the various emission rates cited above for all types of vehicles to be operated in the Los Angeles Basin in 1977. To compute hot- and cold-start emission rates and the emissions/driving speed correlations, a recent study of light-duty vehicle emissions carried out by Automotive Environmental Systems, Inc. (20, 21) can be used. In this study, 1000 motor vehicles from six cities, including Den-
ver, Chicago, Houston, St. Louis, Washington, and Los Angeles, were tested for emissions of NO,, CO, COz, and hydrocarbons using the following test procedures: (1) 1972 Federal Test Procedure (FTP), (2) 1975 FTP, and (3) steady cruise operation at idle, 15, 30, 45, and 60 mph. Since only 1957-71 model year vehicles were tested by AESi, other means must be developed to estimate emissions from 1972-77 vehicles. The assumptions made and algorithms employed will be discussed later in this section. Average hot-start, cold-start, or steady cruise emission rates may be obtained by averaging the emissions from each type of motor vehicle using the following equation
(1A) i=l
j=65
where 1, i, and j are indices denoting chemical species, vehicle class, and model year, respectively, and El = average emission rate of species 1 (g/mi) p L = fraction of motor vehicles belonging to class i x,, = fraction of class i vehicles manufactured in model year j m,, = annual or daily mileage (mi/year or mi/day) for class i, model year j vehicles el,, = emission rate of species 1 from class i, model year j vehicles (g/mi) The eight classes of motor vehicles considered in this study are listed in Table A-I. Vehicles are further grouped according to model year, designated by the index j , where j = 65, 66, . . . , 77 corresponds to pre-1966, 1966, . . . , 1977 model years, respectively. The appropriate values of el,, to be cruise emissioni3, the choice depending on the type of average emission factor desired. Volume 9,Number 5, May 1975
441
Distribution of Vehicles by Class a n d Model Year. The statewide distribution of motor vehicles, by model year, for class 1-6 vehicles on July 1, 1972, was obtained from the State of California ARB (22). If we assume that these data will also apply on July 1, 1977, and also that the statewide distribution is applicable in the Los Angeles Basin, values of x l l used in Equation 1-A are given in Table A-11. The age distribution of motorcycles is assumed to be uniform since emissions data as a function of model year are not available. Thus, x l J = 0.0769 (Le,, 1.0/13) for 6 5 i 5 8. [A more detailed age distribution for motorcycles, as of 1971, is given in the 1972 California Statistical Abstract (23) published by the State of California.] The distribution of motor vehicles by class, F ~ is , best determined directly from vehicle registration statistics for the area of interest. Unfortunately, the appropriate data for the Basin were not readily available. Thus, the class distribution was estimated in the following manner. First, 1972 vehicle registration data for Los Angeles and Orange Counties listing the total number of automobiles, trucks, and motorcycles were used to calculate the relative distribution of these three types of vehicles. Statewide registration data (22) for each class of light- and heavy-duty vehicle were then employed to further refine the automobile, truck, and motorcycle distribution cited above. The resulting distribution for the eight classes of vehicles is given in Table A-III.
Table A-I. Motor Vehicle Class Definitions Class
Type of vehicle
1 2 3 a 5 6 7 8
Light-duty domestic automobiles Light-duty foreign automobiles Light-duty domestic trucks Light-duty foreign trucks Heavy-duty gasoline trucks Heavy-duty d iese I trucks Two-stroke motorcycles Four4roke motorcycles
Table A-11, Estimated Distribution of Automobiles and Trucks, by Model Year, on July 1, 1977" Class
1
2
3
4
5 and 6
0.0727 0.0750 0.0825 0.0904 0.0880 0.0775 0.0868 0.0934 0.0794 0.0653 0.0513 0.0313 0.1064
0.0941 0.1712 0.1229 0.1393 0.0950 0.0802 0.0614 0.0472 0.0395 0.0323 0.0231 0.0197 0.0741
0.0757 0.0765 0.0814 0.0853 0.0675 0.0553 0.0637 0.0693 0.0633 0.0522 0.0414 0.0295 0.2390
0.1658 0.2414 0.1725 0.1086 0.0591 0.0539 0.0390 0.0268 0.0228 0.0213 0.0112 0.0123 0.0653
0.0762 0.0802 0.0812 0.0856 0.0684 0.0571 0.0642 0.0688 0.0623 0.0529 0.0425 0.0309 0.2288
Model year
1977 1976 1975 1974 1973 1972 1971 1970 1969 1968 1967 1966
Pre-1966
a T h e data presented in, this Table were derived f r o m t h e California statewide vehicle population o n July 1, 1972. Source: State of California A R B (22).
Annual Mileage. To properly compute an average emission rate, the emissions from each vehicle must be weighted with respect to the annual (or daily) mileage it will accumulate. This weighting is important because late model year vehicles may be driven twice the annual mileage as that of older vehicles during a given year. Further, heavy-duty vehicles tend to accumulate more mileage than light-duty vehicles, and thus their contribution to the average emission rate must be weighted accordingly. Annual mileages for automobiles and trucks as a function of vehicle age were obtained from the State of California ARB (22). These data, adapted for use in 1977, are presented in Table A-IV. Exhaust Emissions. This section is devoted to the presentation of estimated values of e l l , for use in Equation 1-A. An attempt has been made to base all estimates on actual measurements whenever possible, but unfortunately, several emission rates must be determined from computations based on assumptions whose validity cannot be completely verified. Four classes of light-duty vehicles are defined in Equation l-A, including domestic and foreign automobiles and trucks. Assuming that the Federal Driving Cycle is a reasonable representation of the driving conditions encountered on surface streets in the Basin, hot- and cold-start emission rates for 1957-71 light-duty vehicles may be estimated from the AESi (21) study. An emissions/driving speed correlation for computing emissions from vehicles operated on freeways may be derived from the steady cruise emissions at 15, 30, 45, and 60 mph also reported in this study. Due to the limited number of foreign automobiles tested and the fact that no light-duty trucks were included in the Los Angeles vehicle sample, a single emission factor for each model year is calculated for all class 1-4 vehicles by averaging the available test data for that year. As additional measurements become available for class 2-4 vehicles, then separate emission rates may be determined for each of the four classes. The four classes have been treated distinctly, rather than forming one large class of light-duty vehicles, for two reasons. First, the federal emission standards for post-1974 light-duty trucks are not so strict as the standards for automobiles, and second, different exhaust control retrofit devices are scheduled for installation on foreign and domestic vehicles. Thus, as a matter of computational convenience, four classes of lightduty vehicles are treated in this study. The cold- and hot-start emission rates for pre-1972 vehicles for HC, CO, and NO, calculated from a computer generated summary of all AESi data taken in Los Angeles (24) are summarized in Table A-V. Steady-cruise emissions at 15, 30, 45, and 60 mph may be found in the AESi (21) report. Emissions from 1972-77 vehicles were obtained from calculations based on consideration of existing emission standards (as of August 1973) and measured emission characteristics of 1971 model year vehicles [as derived from the AESi (22) study]. (At the time of this study, implementation of the original 1975 HC and CO and 1976 NO, emission standards was delayed for one year. Subsequently, a delay of an additional year was al-
Table A-Ill. Estimated Distribution of Vehicles, by Class, on July 1, 1977 Class
442
Class
1
2
3
4
5
6
7
8
Fraction of total vehicle population
0.6851
0.1605
0.0876
0.0093
0.0162
0.0052
0.0137
0.0224
Environmental Science & Technology
lowed, as noted in the Introduction.) The details of the algorithms employed are too lengthy to include here, but the results are summarized in Table A-V. For a more complete discussion of how these estimates in Table A-V were derived, the reader is referred to Reynolds (7). Exhaust control device deterioration must be considered for HC and CO emissions from post-1965 vehicles and NO, emissions from p o s t - 1 9 6 9 model year vehicles. Deterioration factors for each model year as a function of the age of the vehicle have been compiled by EPA (2). However, these factors are for use with low-mileage emissions. Since most vehicles tested by AESi were not a t low mileage, the EPA deterioration factors must be modified to account for the deterioration present a t the time of the AESi study. The appropriate deterioration factors, dl,, which must be multiplied by the emissions in Table A-V to yield the final values of elrJfor use in Equation 1-A are given by
where d l , (1977) = deterioration factor for model year j vehicles on July l, 1977 d l , (1972) = deterioration factor for model y e a r j vehicles a t the time of the AESi study Calculated values of d l , for pre-1972 light-duty vehicles applicable on July 1, 1977, are summarized in Table A-VI. Two classes of heavy-duty vehicles are considered in this study, including both gasoline- and diesel-fueled trucks. Emissions for gasoline trucks have been estimated by EPA (2),but no attempt has been made to distinguish between hot- and cold-start truck emissions. Thus, the hot- and cold-start gasoline truck emissions used in this study are both equal to the emission rates given in the report cited above. In addition, the appropriate degradation factors are also included in the EPA report. Since emissions for diesel vehicles as a function of model year are not given in the EPA (2) report, values of these parameters were obtained from the State of California ARB (22). The estimated emissions, including degradation, as of July 1, 1977, are presented in Table A-VII. As with gasoline trucks, the entries in Table A-VI1 are used in both hot- and cold-start emission calculations. Emissions from two- and four-stroke motorcycles are summarized by EPA ( 2 ) . Since emissions are not given as a function of model year, it is assumed that all motorcycles of a particular type emit pollutants at the same rate. By employing Equation 1-A and the data either presented or alluded to above, average hot- and cold-start emission rates, as well as steady cruise emissions, have been calculated for Los Angeles as of July 1, 1977, and are presented in Table A-VIII. Further details with regard to the treatment of steady cruise emissions may be found in Reynolds ( 7 ) . ai and bl are constants to be determined from the values of El(u) given in Table A-VIII. If the given correlation is reasonable, then a log-log plot of emissions vs. driving speed should yield a straight line. It can be shown that NO, emissions are correlated well over the entire driving speed range. HC and CO emissions, however, exhibit marked deviation from linear behavior a t the 1 5 and 60 mph data points, respectively. While emissions for these species generally lie on a line for three consecutive speeds, the correlation does not seem to hold over the entire driving speed range. In an effort to minimize the discrepancy between the correlation function and the predicted emissions, two sets of “constants” are given for HC and CO, where the particular set to be used in a par-
Table A-IV. Estimated Annual Mileage to Be Accumulated by Motor Vehicles in 1977, Mi/Yeara Class Model year
land3
2and4
5
6
1977 1976 1975 1974 1973 1972 1971 1970 1969 1968 1967 1966 Pre-1966
14,400 14,100 12,900 11,400 8,600 6,800 5,700 4,800 4,000 3.600
12,500 12,500 11,600 11,400 10,300 9,500 8,400 5,900 4,000 3.100 2 400 2; 100 2,000
19,600 19,600 18,000 18,000 14,000 14,000 11,000 11,000 8,400 8.400 4:300 4:300 4,300
28,140 28,140 25,800 25,800 20,110 20,110 15,870 15,870 12,140 12.140 6 130
:
3:500 3;500 3,500
:
6:iSo
6,130
.Source: State of California ARB (22). Diesel mileages based on estimated diesel fuel consumption data.
Table A-V. Estimated Cold- and Hot-Start Emission Rates for Light-Duty Vehicles in Los Angeles in 1977, G/Mia Model year
Cold-start HC
19776 0.42 1977~ 2.17 1976b 0.42 197& 2.17 1975* 0.73 1975~ 2.17 1974 2.17 1973 2.17 1972 2.43 1971 3.51 1970 5.22 1969 5.87 1968d 5.36 1967 6.22 1966 8.73 Pre-1966 11.81
Hot-start
CO
NO,
HC
CO
NO,
3.4 18.6 3.4 18.6 5.3 18.6 18.6 18.6 22.9 51.9 62.6 87.1 75:3 81.4 78.1 106.3
0.31 1.49 0.60 1.49 1.49 1.49 1.49 1.90 2.31 3.82 4.51 5.45 3.90 3.30 3.24 2.94
0.08 0.73 0.08 0.73 0.14 0.73 0.73 1.03 1.48 2.65 3.77 4.26 3.84 4.60 7.10 10.02
0.59 5.4 0.59 5.4 0.91 5.4 5.4 7.7 12.2 35.0 41.6 53.3 46.2 55.2 54.2 87.5
0.31 1.49 0.60 1.49 1.49 1.49 i.49 1.90 2.31 3.89 4.68 5.39 4.13 3.50 3.40 2.91
vehicles are subject t o exhaust control device deterioration (see text). *Emissions for automobiles Emissions for trucks. d,Emissions from vehicle number 53 in t h e AESi (1973) data base are n o t included.
Table A-VI. Deterioration Factors for Pre-1972 Light-Duty Vehicles in 1977. Model year HC co NO, 1971 1970 1969 1968 1967 1966 Pre-1966
1.16 1.13 1.10 1.10 1.05 1.04 1.00
1.30 1.17 1.14 1.18 1.16 1.04 1.00 ~~
1.14 1.00 1.00 1.00 1.00 1.00 1.00
“ T h e s e deterioration factors are only suitable t o use with t h e emission rates given in Table VI.
Table A-VII. Estimated Heavy-Duty Diesel Truck Emissions on July 1, 1977. (G/mi) ~
Model year
Pre-1973 73 74 75 76 77 a
HC
co
NO,
4.4 2.4 2.8 1.2 1.0 0.8
26.0 15.0 15.0 12.0 11.0 9.0
43.0 46.0 45.0 15.0 15.0 13.0
Source: State of California ARB (22).
Volume 9, Number 5, May 1975
443
Table A-VIII. Estimated Average Emission Rates for Motor Vehicles in Los Angeles as of July 1, 1977 (G/mi) Test
HC
co
NO,
Cold-start Hot-start
3.20 2.14 1.66 1.14 1.04 0.95
35.0 21.5 24.0 9.9 7.0 7.9
2.58 2.60 0.32 1.13 2.64 3.83
15 r n p h 30 rnph 45 m p h 60 r n p h
Stead cruise emissions are related t o t h e vehicle speed through use of t i e following correlation €dV)
= ,i(V)*I
where €i(v) = emission rate of species 2 a t speed speed, mph.
v = d l ,iving
V
Table A-IX. Emissions-Driving Speed Correlation Parameters for Motor Vehicles in Los Angeles in 1977 Drivin
speeB
Species
range, mph
ai
bi
HC HC
15-30 30-60 15-45 45-60 15-60
7.253 2.746 4.288 X lo2 1.498 2.094
-0.544 -0.259 -1.080 0.406 1.846
co co
NO,
Table A-X. Fuel Tank and Carburetor Evaporative Losses from 1957-71 Light-Duty Vehicles. Evaporative losses (grams per test)
a
Model year
Tank
Carburetor
1957-69 1970-71
25.8 16.1
14.6 11.0
Source: AESi (97).
Table A-XI. Total Daily Motor Vehicle Evaporative Emissions, G/Day
a
Model year
Light-dutya
Heavy-dutya
Pre-1970 1970-71 1972 Post-1972
74.4' 52.7 2.0 2.0
74.4 74.4 74.4 2.0
Gasoline-powered vehicles.
ticular situation will depend on the driving speed. The computed values of al and bl are given in Table A-IX. Daily Vehicle Miles Traveled. The daily vehicle miles traveled (VMT) for both freeways and surface streets in 1969 delineated on a 2-mile grid have been given by Roth et al. (16). Since the population in several areas of the modeling region is expected to grow, the VMT in these areas is also expected to show a similar increase. Liu (17) has estimated that the VMT in the Basin will increase by about 17% between 1969 and 1977. The VMT increase in each 2 x 2 mi grid square can then be computed by apportioning the 17% increase in total mileage according to projected population growth in each square (7). The temporal distributions of surface street and freeway traffic in 1977 are assumed to be identical to those derived for 1969. Similarly, the fraction of cold starts to total starts and the correction for the nonuniform distribution of cold starts are also taken as unchanged from the 1969 values (16). Evaporative Emissions. Evaporative controls on 1972 and new cars will reduce emissions substantially, although about 50% of the vehicle population will be uncontrolled in 1977. Evaporative emissions from 1957-71 light-duty vehicles can be estimated from tests of 125 automobiles performed by Automotive Environmental Systems, Inc. (21). Average emission factors for fuel tank and carburetor losses determined by AESi are summarized in Table A-X. Total daily evaporative losses from each vehicle may be estimated by combining the tank and carburetor losses in the following manner. Total daily carburetor losses are obtained by multiplying the carburetor emissions by the number of daily hot-soaks. From Roth et al. (16) it was estimated that the average vehicle makes 4.66 trips per day, and of these, two are cold-started. Assuming that only half of the hot-soak carburetor losses occur when a vehicle is restarted during a hot start, then the total daily carburetor losses are obtained by multiplying the test re0.5 x 2.66). Total sults in Table A-X by 3.33 (i.e., 2 daily evaporative emissions from each vehicle are obtained by adding the fuel tank losses to the total carburetor losses. The results of this calculation are summarized in Table A-XI. Note that the value of 74.4 g/day for pre-1970 light-duty vehicles agrees closely with the value of 72 g/day used in the 1969 validation study (16). Evaporative emissions for 1973 motor vehicles ( 4 ) appear to be quite small, and thus, post-1972 light- and heavy-duty vehicle emissions are assumed to be 2 g/day, and pre-1973 heavy-duty truck emissions are assumed equal to those of uncontrolled light-duty vehicles. Motorcycle losses are calculated from EPA (2) emission estimates, assuming that each motorcycle travels 3900 miles per year, using the following relationship:
+
0.3 6(g/mi) x 3 9 0 0 ( m i / y e a r ) 365 ( d a y s / y e a r ) = 3.8 (g/day) Table A-XII. Projected Power Plant Emissions in 1977, Tons/Day Power plant
Alarnitos El Segundo Redondo Beach Huntington Beach Long Beach Harbor Haynes Scattergood Valley Pasadena Burbank
Glendale
444
Environmental Science 8, Technology
NOz (as NO?)
27.0 12.2 18.0 15.9 4.1 6.1 37.6 18.7 9.4 2.6 3.6 3.7
Emissions from heavy-duty diesel vehicles are assumed to be negligible. To obtain the total daily evaporative emissions emitted into the airshed, the emission rates given in Table A-XI must be multiplied by the total number of vehicles of each class and model year. If we use 1972 vehicle registration statistics (23) and projected registration growth (25), it is estimated that in 1977 there will be 4,919,700 automobiles, 688,800 trucks, and 210,000 motorcycles in the region under consideration. The total number of vehicles in a particular class and model year grouping is obtained from the projected number of vehicles cited above and from the distribution of motor vehicles, both by class and model year. Total daily evaporative emissions in 1977 are
Table A-XIII. Summary of Assumptions of Effectiveness of Motor Vehicle Emission Controls in 1977 EPA Control Strategy Control program
Assumptions
The VSAD device is scheduled for installation On all 1957-65 light-duty vehicles with engines larger t h a n 140 in.3 u p o n change of ownership a n d on all 1966-70 light-duty vehicles (1957-65 foreign light-duty vehicles assumed t o b e smaller t h a n 140 in.9 The percentage emission reductions are given i n Table A-XVIII. I f we assume t h a t 80% of t h e 1957-65 vehicles will have changed ownership by 1977 (I), t h e overall percentage reduction i n emissions from 1957-65 domestic vehicles is obtained b y multiplying t h e values i n Table A-XVI b y 0.8 To reduce HC a n d CO emissions, t h e installation of oxidizing catalyst retrofit devices will b e required on those 1966-74 light-duty vehicles capable of r u n n i n g on lead-free gasoline. Holmes et al. (30) have estimated t h a t 20% o f t h e pre1971 a n d 75% of t h e 1971-74 model year vehicles will b e eligible for these devices. HC a n d CO reductions average 50%, while NO, emissions are unaffected
VSAD retrofit
Oxidizing catalyst retrofit
(25)
Although t h e exact nature of t h e inspection has not yet been specified, i t will probably b e similar t o either t h e idle or key m o d e tests (31). I n tests comprised of subjecting a fleet of vehicles t o mandatory inspectionmaintenance procedures, it was f o u n d t h a t HC, CO, a n d NO, emissions were reduced b y 12, 10, a n d O%, respectively (25) Limitation of t h e total n u m b e r of registered motorcycles t o t h e 1973 level would reduce t h e 1977 baseline motorcycle emissions by about 9%. I n addition, twostroke motorcycles will not b e allowed t o operate between t h e hours of 6 and 8 p.m. from May t o October To discourage use of low-occupancy motor vehicles during peak traffic hours, one lane on all three-lane freeways a n d t w o lanes on all four-lane freeways would b e reserved for t h e exclusive use of buses and carpools (vehicles containing three or more persons) between t h e hours of 6:30 a n d 9:30 a.m., a n d 3:30 a n d 6:30p.m. I n addition, one lane on all three-lane surface streets will also b e reserved for buses a n d carpools d u r i n g these hours. The following assumptions have been employed i n this study t o estimate the VMT reduction associated with exclusive bus/carpool lanes. First, the VMT associated with freeway traffic traveling i n t h e “slow” direction i s assumed t o b e reduced b y 20% d u r i n g the hours 6:30-9:30 a.m. a n d 3:30-6:30 p.m. This percentage reduction could b e accomplished i f (1) one t h i r d o f t h e people traveling in t h e slow direction form carpools or (2) 15% of the people originally traveling i n t h e slow direction take buses a n d 10% form carpools. Case 1simply states t h a t everyone traveling i n one lane of a three-lane freeway forms a carpool. The traffic congestion i n t h e two regular lanes would b e n o greater t h a n t h a t which would normally occurs. Congestion in t h e exclusive lane would b e less t h a n usual, resulting i n faster driving speeds. Since freeway driving speeds i n the “fast” direction are generally high (16),t h e exclusive lanes are assumed t o have no effect o n t h e VMT associated with freeway travel i n t h e “fast” direction. The formation of exclusive lanes on surface streets is assumed t o have a negligible effect on surface street VMT. However, surface street VMT would b e reduced t o some extent b y those people causing t h e assumed 20% VMT reduction on freeways cited above. For an average home-to-work trip, the ratio of surface street miles t o freeway miles is 1.75 (25). I f n o VMT were required t o form carpools or get people t o buses, t h e n t h e total surface street VMT savings would b e c o m p u t e d by multiplying t h e freeway mileage savings by 1.75.To account for some VMT d u e t o carpool formation, t h e surface street mileage reduction calculated above is multiplied by 0.75. This savings in surface street VMT is 10% of t h e original 6:30-9:30 a.m. mileage. The m o s t effective means for reducing t h e Basinwide VMT is t h e limitation of gasoline sales. I n t h e EPA strategy, annual gasoline sales will not b e allowed t o exceed t h e total a m o u n t sold between July 1, 1972, and June 30, 1973. To c o m p u t e t h e percentage reduction i n VMT, note t h a t gasoline consumption is increasing b y 4.5% each year a n d t h a t t h e projected total Basinwide VMT increase i s estimated t o b e 17% between 1969 a n d 1977. Assuming t h a t t h e total VMT is proportional t o gasoline consumption, the result of t h e limitation on gasoline sales will b e t o reduce VMT by 15%. Due t o t h e “essential” nature of home-to-work trips t h a t take place between 6:30 a n d 9:30 a.m., it is assumed t h a t t h e only VMT reduction during this three-hour period is caused b y t h e implementation of exclusive bus-carpool lanes. After treating t h e “essential” work-to-home trips during the evening i n a n analogous manner, t h e Basinwide VMT m u s t b e reduced b y 17.5% during t h e remaining hours of t h e day (i.e., excluding the morning and evening hours) t o obtain an overall 15% daily VMT reduction.
Mandatory inspection/ maintenance
Motorcycle registration limitation a n d operation b a n Exclusive bus/carpool lanes
Gasoline sales limitations in 1972-73 levels
Table A-XIV. Hot- and Cold-Start Emission Factors Estimated for 1977 EPA Strategy, G/Mia co
HC
e s t i m a t e d t o b e 204 t o n s o f h y d r o c a r b o n s p e r d a y . T h e treatment o f t h e temporal a n d spatial distribution o f evaDorative emissions i s discussed in R o t h e t a l . (16). . ,
NO, (as NO,)
1977 Baseline Inventory-A ircraft Cold start H o t start
Gas
Diesel
Gas
Diesel
Gas
Diesel
2.09
0.03 0*03
25.4 16’’
0.2 o’2
1.95
0.35 0‘35
“ T h e figures given in this table are t h e relative contributions fr,om gasoline- a n d diesel-powered vehicles t o t h e “average“ motor vehicle emission rate obtained from Equation 1-A.
To e s t i m a t e a i r c r a f t emissions in 1977, t h r e e changes h a v e b e e n m a d e in t h e 1969 a i r c r a f t i n v e n t o r y r e p o r t e d by R o t h e t al. (16), i n c l u d i n g t h e a d d i t i o n o f a n o t h e r class o f aircraft, t h e r e v i s i o n of j e t a i r c r a f t e m i s s i o n factors, a n d t h e p r o j e c t i o n of t h e n u m b e r o f d a i l y a i r c r a f t operations a t e a c h a i r p o r t in 1977. S i n c e w i d e b o d y jets, s u c h as t h e Volume 9, Number 5, May 1975
445
Boeing 747, Lockheed L-1011, and Douglas DC-10, will be operating at Loe Angeles International Airport (LAX) in 1977, an eighth class of aircraft, characterized by the JTSD engine, has been added to the 1969 inventory. Total daily operations in 1977 a t each airport in the modeling region were estimated from Federal Aviation Agency (FAA) projections and from conversations with the managers of several small airports. For further details of the estimation of aircraft emissions in 1977, the reader is referred to Reynolds ( 7 ) . 1977 Baseline Inventory-Fixed
Sources
Since few guidelines are available for estimating future fixed source emissions directly, projections made in this study are based on indirect relationships, such as the assumption that increased motor vehicle fuel consumption will cause refinery emissions to increase proportionately. The three types of stationary sources to be considered in this section are power plants, oil refineries, and other miscellaneous fixed sources. Power plant emissions are affected by the demand for electricity and the type of fuel burned. However, information with regard to future electrical demand and fuel usage at each power plant are generally not available. Fortunately. projected NOx emissions in 1977 at each of the power plants to be operated by the Southern California Edison CO. and the Los Angeles Department of Water and Power were furnished by the respective power companies (26, 2 7 ) . Emissions from power plants operated by the Cities of Pasadena, Burbank, and Glendale were estimated by increasing the reported emissions as of April 1973 ( 2 8 ) by 38% as suggested by Nevitt ( 2 9 ) . The daily NO, emissions projected for each power plant in 1977 are summarized in Table A-XII. The temporal distribution of power plant emissions for 1977 is taken to be the same as that employed in 1969 ( 7 ) . Since the construction of no new oil refineries in the Basin is expected by 1977, only changes in emissions from existing refineries are considered in this study. Due to the lack of guidelines for estimating future refinery emissions, it is assumed that emissions will increase in proportion to increases in gasoline fuel consumption. Assuming as an upper limit a 4.5% annual growth of gasoline consumption, refinery emissions are estimated to increase by about 36% from the corresponding 1969 values. Because all emissions are assumed to increase by the same percentages, no attempt was made to alter the relative spatial distribution of emissions ( 7 ) . Note that all refinery emissions. including RHC, URHC, and NO,, will increase by 36%. Emissions from all fixed sources other than power plants and refineries have been estimated by Liu ( 1 7 ) . These sources include such items as industrial and domestic boilers, petroleum marketing operations, and organic solvent usage. If we use the work of Trijonis ( 1 5 ) as a guide, and assume that emissions resulting from petroleum marketing operations are proportional to fuel consumption, the 1969 RHC and NO, emissions are estimated to increase by 23.9% and 28.6%, respectively. URHC emissions are also assumed to increase by 23.9%. 1977 EPA Control Strategy Inventory
In this section the quantitative impact on emissions of the EPA plan is estimated. Motor Vehicle Emissions. Due to the extensive control measures proposed for motor vehicles, all aspects of the 1977 baseline motor vehicle emissions inventory given above must be revised. A major difficulty in estimating motor vehicle emissions is that retrofit device emission reduction factors have not 446
Environmental Science &
Technology
Table A-XV. Emission-Driving Speed Correlation Parameters Estimated for 1977 EPA Control Strategy. Drivana
spezd Polrange, lutant mph
HC HC CO
CO NO,
15-30 30-60 15-45 45-60 15-60
bi
ai
Gas
Diesel
5.206 1.821 2.205 X 10’ 1.279 1.523 X 10V3
9.823 X 4.484 X 6.691 4.645 X 2.761 X
Gas
Diesel
-0.587 -0.278 -0.980 0.372 1.868
-0.553 -0.322 -1.266 0.644 1.855
a The figures in this table a,re for use in calcu!ating the relative contributions of gasoline- and diesel-powered vehicles t o the “average“ freeway emission rates.
Table A-XVI. Percentage VSAD Emission Reductions for Domestic and Foreign Light-Duty Vehicles co
HC
NO,
Model year
Dom.
For.
Dom.
For.
Dom.
For.
1957-65 1966-70
254 23b
0 3c
95 6b
0 3
230 44b
42c
0
a Source: Horowitz (33)’ emission reductions also subject t o change of ownership factor (see tekt). Source: Lees et al. (39). e Source: State of California A R B ( 3 4 ) .
been reported for all types of driving conditions (that is, reductions based on hot-start, cold-start, and steady cruise tests). For lack of better information, it is assumed here that reported emission reductions for a particular device apply to all driving modes. The effectiveness of exclusive bus/carpool lanes on reducing VMT is also largely unknown. Thus, the emission reductions estimated here are, at best, speculative and subject to revision as better information becomes available. Table A-XIII summarizes the assumptions made in regard to the effectiveness of each element of the plan designed to reduce motor vehicle emissions. To estimate average motor vehicle emission factors, the values in Table A-V must be reduced based on the cumulative effectiveness of the VSAD, oxidizing catalyst, and inspection/ maintenance programs. The contributions from gasoline- and diesel-powered vehicles to the “average” hot- and cold-start emission factors are given in Table A-XIV. Diesel vehicle emissions are calculated by multiplying the factors in Table A-XIV by the 1977 baseline surface mileage in each grid square; gasoline vehicle emissions are calculated by multiplying the factors in Table A-XIV by the adjusted 1977 baseline mileages (i.e., modified to reflect the percentage VMT reductions from the exclusive lane and gasoline sales limitation strategies). This treatment ensures that diesel emissions will not be affected by strategies aimed at reducing VMT associated with gasoline-powered vehicles. Freeway emissions are treated analogously; the emissions-driving speed correlation parameters for both gasoline and diesel vehicles are summarized in Table A-XV. Daily evaporative losses from motor vehicles are estimated as follows: Emissions from the fuel tank are assumed to be unaffected by the control strategies, but carburetor losses are reduced by 15%, corresponding to the basinwide VMT reduction. This is analogous to assuming that the total daily trips, and hence carburetor losses, are also reduced by 15%. The total evaporative emissions are estimated to be 184 tons per day. Aircraft Emissions. The EPA control strategy is as-
sumed to affect neither the aircraft emission factors nor the flight operation data employed in the baseline inventory. Fixed Source Emissions. Although specific control measures for fixed sources were not proposed in the control plan, EPA is currently considering such measures as installation of vapor collection and disposal systems and substitution of nonreactive solvents for the reactive solvents presently employed by organic solvent users. Employing the work of Trijonis (15) as a guide, Liu (17) has estimated that implementation of the EPA control strategy will reduce RHC and NO, emissions from oil refineries by 58% and 6170, respectively, and that RHC and NO, emissions from other fixed sources will be reduced by 81% and 5470, respectively. These percentage reductions are with respect to the 1977 baseline inventory discussed above. Power plant emissions were assumed to be unaffected by the EPA strategies.
Literature Cited (1) Environmental Protection Agency, “Technical Support Document for the Metropolitan Los Angeles Intrastate Air Quality Control Region, San Francisco, California,” 1973. (2) Environmental Protection Agency, “Compilation of Air Pollutant Emission Factors,” 2nd ed., Rep. AP-42, Off. Air and Water Progr., Research Triangle Park, N.C., 1973. (3) Environmental Protection Agency, “Federal Certification Test Results for 1973 Model Year,” Fed. Regist., 38, 10868 (1973). (4) Environmental Protection Agency, “California Air Quality Standards-Approval and Promulgation of Implementation Plans,” ibid., p 2194. (5) Reynolds, S. D., Roth, P. M., Seinfeld, cJ. H., “Mathematical Modeling of Photochemical Air Pollution. I. Formulation of Mode1,”Atmos. Enuiron., 7, 1033 (1973). (6) Reynolds, S. D., Liu, M., Hecht, T . A., Roth, P. M., Seinfeld, J . H., “Mathematical Modeling of Photochemical Air Pollution. 111. Evaluation of Model,” ibid , 8, 563 (1974). (7) Reynolds, S. D . , “Mathematical Modeling of Photochemical Air Pollution,” PhD thesis, Cal. Inst. Tech., Pasadena, Calif. (1974). (8) Pierrard, J . M., Snee, R. D., Zelson, J., “A New Approach to Setting Vehicle Emission Standards,” Paper 73-75, 66th Ann. Meet. Air Pollut. Contr. Assoc., Chicago, Ill., (1973). (9) Panel on Emission Standards, “A Critique of 1975 Federal Automobile Emission Standard for Carbon Monoxide,” Nat. Acad. Sci., Washington, D.C. (1973). (10) Chang, T . Y., Gratch, S., Weinstock, B., “The Relationship of Vehicle Emission Standards and Air Quality Standards,” Ford Motor Co., Dearborn, Mich., 1974. (11) Dabberdt, W. F., Ludwig, F . L., Johnson, Jr., W. B., “Validation and Applications of Urban Diffusion Model for Vehicular Pollutants,” Atmos. Environ., 7, 603 (1973). (12) Schuck, E . A,, Papetti, R. A,, “Examination of the Photochemical Air Pollution Problem in the Southern California Area,” Environmental Protection Agency, Washington, D.C., 1973. (13) Dimitriades, B., “Effects of Hydrocarbon and Nitrogen Oxides on Photochemical Smog Formation,” Environ. Sci. Technol., 6, 253 (1972).
(14) Dimitriades, B., “Photochemical Oxidants,” Chemistry and Physics Laboratory, Environmental Protection Agency, Research Triangle Park, N.C., 1973. (15) Trijonis, J. C., “An Economic Air Pollution Control Model-Application: Photochemical Smog in Los Angeles County in 1975,” PhD thesis, Cal. Inst. Tech., Pasadena, Calif., 1972. (16) Roth, P. M., Roberts, P. J. W., Liu, M., Reynolds, S. D., Seinfeld, J . H., “Mathematical Modeling of Photochemical Air Pollution. 11. A Model and Inventory of Pollutant Emissions,” Atrnos. Enuiron., 8,97 (1974). (17) Liu, M. K., private communication, Systems Applications, Inc., San Rafael, Calif., 1973. (18) Panels on Emission Standards and Atmospheric Chemistry, “A Critique of the 1975-76 Federal Automobile Emissions Standards for Hydrocarbons and Oxides of Nitrogen,” Nat. Acad. Sci., Washington, D.C., 1973. (19) Dimitriades, B., personal communication, Environmental Protection Agency, 1974. (20) Fegraus, C. E., Domke, C. J., Marzen, J., “Contribution of Motor Vehicle Population to Atmospheric Pollution,” SOC. Auto. Eng. Paper No. 730530, Presented May 14-18, Auto. Eng. Meet., Detroit, Mich., 1973. (21) Automotive Environmental Systems, “A Study of Emissions from Light-Duty Vehicles in Six Cities,” U S . Environmental Protection Agency Rep. APTD-1497, Westminster, Calif., 1973. (22) Bratton, D., personal communication, State of California Air Resources Board, Sacramento, Calif., 1973. (23) County Supervisors Association of California, “California County Fact Book,” Sacramento, Calif., 1972. (24) Sachtsdale, J. R., personal communication, Automotive Environmental Systems, Westminster, Calif., 1973. (25) TRW, “Transportation Control Strategy Development for Metropolitan Los Angeles Region,” Rep. APTD-1372, Redondo Beach, Calif., 1973. (26) Felgar, D. N., private communication, Southern California Edison Co., Rosemead, Calif., 1973. (27) . , Sonderlinp.. H.. ibid.. Los Angeles De& Water and Power. Los Angeles,?alif:, 1973: (28) Los Angeles County Air Pollution Control District, “Major Point Sources of Air Pollution in Los Anneles Countv-Ami1 1973,” Los Angeles, Calif., 1973. (29) Nevitt, J . S., personal communication, Los Angeles Air Pollution Control District, Los Angeles, Calif., 1973. (30) Holmes, J., Horowitz, J., Reid, R., Stolpman, P., “The Clean Air Act and Transportation Controls’’-an EPA White Paper, U.S. Environmental Protection Agency, Off. Air and Water Progr., Washington, D.C., 1973. (31) Horowitz, J., “Inspection and Maintenance for Reducing Automobile Emissions,” J . Air Pollut Contr. Assoc., 23, 273 (1973). (32) Lees, L. et al., “Smog: A Report to the People,” Environmental Quality Laboratory, Calif. Inst. Tech., Pasadena, Calif., 1972. (33) Horowitz, J., “The Effectiveness and Cost of Retrofit for Reducing Automobile Emissions,” J . Air Pollut. Contr. Assoc., 23, 395 (1973). (34) State of California Air Resources Board, “STP Corporation Application and Accreditation of a Modified NO, Control Device for Certain 1966-70 Model-Year Vehicles with Engines in Class (a) and in Classes ( b j through (f),” Staff Rep. No. 73-10-2, Sacramento., Calif., 1973. L
L
I
.
Received for review March 8, 1974. Accepted December 12, 1974. Work supported by Enuironmental Protection Agency Contract 68-02-0339 to Systems Applications, Znc., by National Science Foundation Grant ENG 71-02486and agift of the John A . McCarthy Foundation to the California Institute of Technology
Volume 9, Number 5, May 1975
447
