EPA-AA SDSB 79-30 Technical Report An Investigation of Future Ambient Diesel Particulate Levels Occurring In Large-Scale Urban Areas By Daniel P. Reiser November, 1979 NOTICE Technical Reports do not necessarily represent final EPA decisions or positions. They are intended to present technical analysis of issues using data which are currently available. The purpose in the release of such reports is to facilitate the exchange of technical information and to inform the public of technical devel- opments which may form the basis for a final EPA decision, position or regulatory action. Standards Development and Support Branch Emission Control Technology Division Office of Mobile Source Air Pollution Control Office of Air, Noise and Radiation U.S. Environmental Protection Agency ------- -2- I. Introduction EPA proposed a particulate standard for diesel-powered light- duty vehicles in February, 1979, and is in the process of promul- gating this standard. One of the prime inputs to this process is the effect of diesel particulate emissions on air quality. The purpose of this report is to determine the diesel's effect on ambient particulate levels over large urban areas. Past studies will be examined and combined with original projections to arrive at the best estimate of ambient diesel particulate levels in U.S. cities. A companion study is being conducted to determine the diesel's impact in smaller, local areas where the impact may be significantly larger (e.g., street canyons). The original calculations of future regional impacts from diesel particulate emissions will be based on past ambient lead measurements in various urban areas. Almost all of ambient lead prior to 1976 can be traced to lead-containing particulate from automobile exhaust emissions. By relating future particulate emissions from diesels to past lead emissions from gasoline-fueled vehicles, ambient lead concentrations of the past can be used to project ambient diesel particulate concentrations in the future. General Motors (GM), in their comments to the proposed light- duty diesel particulate standard, did just this, they projected ambient levels of diesel particulate for the year 1990 using ambient lead concentrations found in two major cities, Toledo and Chicago. However, GM did not document their methodology to support their calculations. Since it is desirable to project the ambient impact of diesel particulate in as many cities as possible, GM's work will be repeated to confirm their results and then expanded to other cities using a documented methodology. The results from this work will then be compared to two studies performed by PEDCo Environmental examining the impact of diesel particulate emissions in 1) Kansas City and 2) New York, Chicago, and Los Angeles. The rest of this report has been divided into five sections. The first section contains the development of a reasonable scenario which includes future diesel sales, future traffic levels, and diesel particulate emission factors. The second section contains 1) a survey of the three studies which have already examined the air quality impact of the diesel and 2) modifies the results of these studies to conform with the scenario developed in the pre- vious section. In the third section, the lead surrogate approach (GM) is outlined and extended to many other cities. The fourth section contains a comparison of the results of the three studies (including the extention) while the final, fifth section contains the conclusions of the analysis. II. Development of Scenario Whenever two studies predicting the same phenomena are com- ------- -3- pared, differences can result because of two factors. One, the input data may differ. Two, the methodology may differ. This study is primarily concerned with differences in methodology. The study's goal is to arrive at the best estimate of the ambient impact of a specified level of diesel particulate emissions. In order to compare methodologies from one study to another the same input data must be used in every case. This input scenario will consist of projected emission factors, growth rates for vehicular traffic and the breakdown of this traffic by vehicle class. EPA estimates that by 1990 the uncontrolled particulate emission factor from light-duty diesels will be 1.0 gram per mile (g/mi).J_/ This emission factor anticipates an increase in particu- late exhaust emission due to a more stringent NOx emission standard being implemented by 1985. The heavy-duty particulate emission factor is presently estimated to be 2.0 g/mi.^/ It will be assumed that these emission factors would not change by 1990 without regulation. A summary of emission factors corresponding to vehicle class is shown in Table 1. Because this report is only concerned with diesel particulate emissions, particulate emissions from future gasoline-fueled vehicles will be assumed to be zero. The breakdown by vehicle class of total urban vehicle miles traveled (VMT) by 1990 is also shown in Table 1. Two different breakdowns are shown: a "low estimate" and a "high estimate" based on a range of projected diesel sales. When engine type (gasoline or diesel) is ignored, the breakdown by class is the same in both cases and was based on DOT data.2/ EPA's latest projection of diesel penetration into the 1990 light-duty vehicle and truck fleet was the basis for the breakdown of the first two classes of Table 1._3/ That projection only included a single best estimate. To indicate the error possible in such projections, a range of plus and minus 25 percent of the best estimate was used. The estimate of diesel penetration into the heavy-duty fleet was taken from the Regulatory Analysis for EPA's proposed light-duty diesel particu- late regulations.^/ Knowing the emission factors and the fraction of VMT for each class, an average weighted emission factor of 0.17 g/mi for the "low estimate" case and 0.27 g/mi for the "high estimate" case can be calculated. These figures were obtained by summing the product of each urban VMT fraction and its correspon- ding emission factor. From examining the contributions of each class to these average emission factors it can be seen that light- duty diesels contribute 56.7 percent and 60.6 percent of total diesel particulate emissions, low and high estimates, respec- tively. Finally, an estimate of urban traffic growth is necessary to obtain a complete picture of future vehicle emissions. An annual growth rate of 1 percent will be used here. This should be suit- able as it has been EPA's policy to use a 1 percent per year growth rate for projections of CO emissions, which, along with diesel ------- -4- Table 1 Fraction of Urban VMT by Mobile Source Category in 1990 Fraction VMT (1990) Classification Low High LDV-G LDV-D LDT-G LDT-D HDT-G HDT-D 0.745 0.085 0.096 0.012 0.025 0.037 0.689 0.141 0.089 0.019 0.010 0.052 Emission Factor 0 1.0 0 1.0 0 2.0 LDV = Light-duty vehicles. LDT = Light-duty trucks. HDT = Heavy-duty trucks. G = Gasoline. D = Diesel. ------- -5- particulate, is primarily an urban core problem.^/ This increase should be compounded between the last year of traffic or ambient pollutant measurement and the year being examined (1990). In- cluding traffic growth in future projections is necessary as it alone will cause an increase in ambient particulate concentrations, if other factors are left unchanged. III. Survey and Modification of Previous Studies A. PEDCo-Kansas City PEDCo Environmental recently performed a study on the impact of diesel particulate emissions on total suspended particulate (TSP) concentrations in the atmosphere,2j Specifically, the Air Quality Display Model (AQDM) was used to predict the diesel's impact on the air quality of Kansas City, Missouri. The city was broken down into 165 sections using a grid of 2 x 2 kilometer squares. The average diesel particulate concentration in each square was determined. EPA requested and examined PEDCo's grid data sheet for one case and determined the fraction of the pop- ulation exposed (by residence) to various levels of diesel part- iculate in 1990._4/ These results are shown in the second column of Table 2. ~ In calculating the impacts shown in the second column of Table 2, PEDCo used emission factors of 0.5 grams per mile (g/mi) for light-duty diesels and 2.0 g/mi for heavy-duty diesels with a traffic growth rate of 1.51 percent per year between 1974 and 1990. The traffic breakdown (percentage of total regional travel) used was 19.1 percent for light-duty diesel vehicles, 2.4 percent for light-duty diesel trucks, and 5.2 percent for heavy-duty diesel trucks. To convert these results to the scenario of Section II, two converting factors must be determined. One is the ratio of the traffic-weighted diesel particulate emission factors and the other is the ratio of future traffic levels between 1974 and 1990. From the figures shown in the preceding paragraph, PEDCo's weighted emission factor was 0.21 g/mi and their overall estimate of traffic growth was 27 percent. From Section II, EPA's weighted emission factors are 0.17 g/mi (low) and 0.27 g/mi (high) and the estimate of traffic growth is 17.3 percent (16 years). The ratios of emission factors are then 0.81 (0.17/0.12) and 1.29 (0.27/0.21), low and high diesel estimates, respectively. The ratio of future traffic levels is 0.92 (1.173/1.27). Combining these two factors, the PEDCo results need to be multiplied by 0.74 and 1.14 to be converted to EPA's low and high diesel scenarios, respectively. This has been done and the modified results are shown in the last two columns of Table 2. B. PEDCo-New York, Chicago, and Los Angeles PEDCo performed a second study on the environmental impact of ------- Table 2 Predicted Population Exposure to Ambient Particulate Levels for Kansas City, Mo, in 1990 % of to Population Exposed Estimated Ambient Levels 2.1 5.9 13.2 17.8 28.6 32.8 Ambient Part. Level (ug/m^) Diesels (Pedco, "Max Growth") 1.732 1.586 1.48 1.386 1.24 1.20 Ambient Part. Level (ug/m^) Diesels (EPA, "Low Est.") 1.3 1.2 1.1 1.0 0.92 0.89 Ambient Part. Level (ug/m^) Diesels (EPA, "High Est.") 2.0 1.9 1.7 1.6 1.5 1.4 i ON 1 ------- -7- diesel particulate emissions,5/ this time based on ambient total suspended particulate (TSP) data taken at fifteen monitoring sites, five each in New York, Chicago, and Los Angeles. TSP data collec- ted from these SAROAD sites were converted to ambient diesel particulate levels using the approximation that motor vehicles contribute a fixed percentage of the TSP levels in each city. PEDCo estimated that particulate emissions from motor vehicles in 1975 and 1976 contributed about 21 percent of the TSP in New York City, 13 percent of TSP in Los Angeles, and 17 percent of TSP in Chicago .j>y These percentages were the result of analyses examining the amount of elemental lead in ambient TSP measurements. It is uncertain whether these percentages as discussed in the original references, refer to leaded exhaust particulate only or to all particulate emissions associated with automobiles (e.g., tire particulate emissions, reentrained dust, etc.). PEDCo assumed that the percentages referred to leaded exhaust particulate. This uncertainty will be analyzed in the Comparison section of this report (V). Via the above-mentioned percentages, PEDCo was able to deter- mine the motor vehicle contribution to ambient TSP levels in 1975-1976. They then determined future ambient diesel particulate levels by using 1) the ratio of the diesel particulate emission factor to the leaded-gasoline particulate emission factor, 2) the ratio of future diesel traffic to the existing traffic of leaded- gasoline fueled vehicles, and 3) the future overall traffic growth. Through the use of two or more different estimates of the above three factors, PEDCo examined a total of six scenarios. As PEDCo"s work involved the same basic assumption as that used in this report, that the air quality impact of a source is proportional to the emissions of that source, only one of PEDCo's scenarios need be modified to the scenario of Section II. PEDCo's scenario which assumed optimistic growth in diesel use and traffic will be the one examined here (Scenario Tj D2 Ej). The particulate emission factors used were 0.5 g/mi for light-duty diesels and 2.0 g/mi for heavy-duty diesels. The breakdown of traffic by vehicle and the traffic growth rates used were different for each city and are shown in Table 3. The weighted emission factors resulting from these growth rates and individual vehicle emission factors for the year 1990 are also shown in Table 3. The resulting ambient diesel particulate levels at all fifteen sites are shown in Table 4 (first column). These results, modified to the scenario outlined in Section II, are shown in the second column of Table 4. The methodology used to modify these original PEDCo results was the same as that used to modify the results of the previous PEDCo-Kansas City study. The ratios of the weighted emission factors and the future traffic levels were multiplied against the original results to obtain the modified results. These results are substantially higher than the PEDCo-Kansas City predic- tion. An explanation of this will be discussed in Section V. ------- -8- Table 3 Input Parameters of PEDCo Three-City Study 5/ Tl, D2, El Scenario - 1990 New York Chicago Los Angeles Overall Traffic Growth (1976-1990) 10.! Vehicular Traffic Breakdown by Class Light-Duty* Gasoline Diesel 50.7% 40.5% 6.4% 47.8% 42.2% 22.7% 52.2% 41.6% Heavy-Duty** Gasoline Diesel 0.3% 8.4% o.: 4.1 0. 6. Particulate Emission Factors Light-Duty Heavy-Duty Diesel Particulate Weighted Emission Factors 0.5 g/mi 0.5 g/mi 0.5 g/mi 2.0 g/mi 2.0 g/mi 2.0 g/mi 0.37g/mi 0.307g/mi 0.33 g/mi * Light-duty includes autos and taxi's in PEDCo's terminology. ** Heavy-duty includes heavy-duty trucks and buses in PEDCo's terminology. ------- -9- Table 4 Predicted Ambient Particulate Levels for 1990 from Diesels EPA vs Pedco City New York New York New York New York New York Torrance, Los Angeles Long Beach, Los Angeles Los Angeles Pasadena Pasadena Chicago Chicago Chicago Chicago Chicago Site Adress EPA Scenario PEDCo Scenario (ug/m3) T1>D2»E1 (ug/m3) Low High Steinman Hall ,W. 141 St. and Convent Ave. 170 E. 121 St. Central Park Arsonal, 5th Ave., and 64th St. 240 2nd Ave. Pier 42, Morton St. and Hudson River 2330 Carson St. 2655 Pine Ave. 434 S. Pedro 1196 East Walnut Kech Laboratories, Cal. Inst. of Tech. 3500 E. 114 St. 1947 W. Polk 9800 S. Torrence Ave, 538 S. Clark St. 4015 N. Ashland Ave. 20.50 19.78 20.45 24.90 20.72 22.59 46.91 22.91 25.07 23.82 19.64 9.8 9.6 10.0 12.0 10.1 11.0 28.1 13.7 15.0 14.3 11.8 15.1 21.62 19.97 25.62 24.82 10.3 9.5 12.2 11.8 16.0 14.7 18.9 18.7 15.0 15.4 18.7 15.6 17.1 43.6 21.3 23.3 22.1 18.3 ------- -10- C. GM Study In their response to EPA's proposed standards, GM submitted an air quality impact section in which ambient particulate concentra- tions from diesel vehicle emissions were calculated from a lead tracer model.6J A methodology was not presented in this report, but a scenario was given which included a particulate emission rate of 0.2 g/mi with a light-duty vehicle (LDV) fleet of 25 percent diesels for the year 1990. A 1 percent per year traffic growth rate was also used. The GM results are shown in Table 5, as well as the results modified to EPA's scenario. As the GM and EPA traffic growth rates are the same in this case, no adjustment due to this factor was necessary. The GM weighted diesel particulate emission factor was simply 0.05 g/mi (0.25 x 0.2 g/mi), so the only adjustment was converting this to the 0.17 g/mi and 0.27 g/mi weighted emission factors, low and high diesel estimates, respec- tively. The regional annual means determined by GM were based on annual lead measurements in Chicago and Toledo taken in 1970 and 1968, respectively. GM claimed that this lead surrogate method is sensible and straightforward and can be reliably applied to major U.S. cities as well.jji/ Because of this ease and applicability to many urban areas, this lead surrogate work will be extended to more cities below. IV. Extension of Lead Surrogate Work to Other Cities Although GM used the lead surrogate approach to predict future particulate concentrations from light-duty diesels, the exact methodology was not documented. As it would be helpful to extend this work to other cities, a methodology will first be outlined below and then extended using monitoring data similar to that used by GM. This methodology should be very similar to that used by GM and any differences will be examined in Section V. The basic assumption involved in surrogate work of this type is that the ratio of the ambient level to emissions of one pol- lutant (in this case lead) is related to that of another pollutant (in this case diesel particulate). Lead has the advantage of being easily separable from other particulate components and the great majority of it is emitted from motor vehicles. It is much more difficult to distinguish diesel particulate from carbonaceous particulate from other sources. Thus, the relationship between ambient lead concentrations and lead emissions from motor vehicles is first determined from actual measurements of both. Second, this relationship for lead is modified as necessary to represent the same relationship for diesel particulate. Finally, this relation- ship for diesel particulate is coupled with diesel particulate emission data to yield estimates of ambient diesel particulate levels. ------- -11- Table 5 Results and Modification of GM Study 6/ Ambient Diesel Impact Regional Annual Mean (ug/m^) GMEPA Low High Major Cities (Chicago) 3.2 10.9 17.3 Mid-size (Toledo) 0.9 3.1 4.9 ------- -12- Th e first step in this process is to express the ambient levels of both diesel particulate and lead in terms of their respective emissions. These relationships are expressed in the following two equations: C(Pb) = E(Pb) • f(Pb) (1) C(D) = E(D) • f(D) (2) where: C(Pb) = concentration of ambient lead levels from mobile source emissions in a particular urban area. E(Pb) = average motor vehicle emission factor for lead in a particular urban area. f(Pb) = a function which relates lead emissions to ambient lead concentrations (constant for each monitoring site). C(D) = concentration of ambient diesel particulate levels from diesel mobile source vehicles in a particular urban area. E(D) = average motor vehicle emission factor for diesel particulate for a particular urban area. f(D) = a function which relates diesel emissions to ambient diesel particulate levels (constant for each moni- toring site). As can be seen, the ambient levels of both pollutants have been assumed to be proportional to their emission factor. This is a standard assumption when working with one source of a non-reactive pollutant. Here we are working with many individual sources (i.e., vehicles). If the relative distribution of these vehicle through- out the region were changing, then f(Pb) or f(D) could change if the average emission factor changed. However, for the purposes of this report, the relative distribution of vehicles throughout a region will be assumed to remain constant. The overall breakdown of traffic by class and engine type may change and the overall traffic may increase, however, each subsection of the region is assumed to have the same fraction of the region's total traffic as it had when the lead studies were performed. Under this condition, equations (1) and (2) are quite valid. ------- -13- Equation (2) can not be used alone to calculate concentrations of ambient diesel particulate levels since f(D) is unknown. However, if equation (2) is divided by equation (1), and solved for C(D), C(D) then becomes a function of C(Pb) and two factors, one related to emissions and one related to dispersion. It may be possible to determine the ratio f(D)/f(Pb) where it would not have been possible to determine f(D) alone. The equation is shown below: C(D) = E(D) ' f(D) • C(Pb) E(Pb) f(Pb) In the following three subsections, the three factors shown on the right side of equation (3) will be determined. In the fourth subsection, all three will be combined to yield estimates of ambient diesel particulate levels in a large number of cities throughout the U.S. A. Lead and Diesel Particulate Emissions The first factor of equation (3) to be determined will be that relating to emissions, E(D)/E(Pb). The average emission factor for diesel particulate has already been calculated in Section II and is 0.17 g/mi for the low diesel estimate case and 0.27 g/mi for the high diesel estimate case. The average emission factor for lead will be determined below. Lead emission factors for light-duty vehicles (LDV), light- duty trucks (LDT) and heavy-duty vehicles (HDV), can be determined from three pieces of data; 1) the lead content of gasoline, 2) the fraction of the lead entering the engine that is emitted from the exhaust, and 3) the fuel economy of the vehicle. All of these factors will be determined circa 1975, as this is the year of the ambient lead measurements. The (elemental) lead content of gasoline in 1975 was 1.9 grams per gallon. Tj Past studies have found that approximtely 75 percent of the lead in the fuel leaves through the exhaust.^/ The rest is accumulated in the oil sump and exhaust system. The average fuel economy of light-duty vehicles in 1975 was 13.5 miles per gallon and that for trucks was 8.7 miles per gallon (based on DOC data).jy No further breakdown was available on the fuel economy of trucks into EPA's light-duty and heavy-duty categories so this figure will be used as the average fuel economy for these two classes. Combining these figures, the lead emission factor for light- duty vehicles is 0.105 g/mi and 0.164 g/mi for both light-duty trucks and heavy-duty vehicles. These figures and the breakdown of urban traffic in 1974 (assumed applicable in 1975)2/ are shown in ------- -14- Table 6. Combining the figures of Table 6 yields an average 1975 lead emission factor of 0.11 g/mi. Only one step remains before the ratio E(D)/E(Pb) can be determined. The average diesel particulate emission factor (0.17 and 0.27 g/mi) is in terms of 1990 miles, while the lead emission factor is in terms of 1975 miles. Between these two years, however, overall travel will increase by 16.1 percent (1.0 percent annual growth compounded for 15 years). Thus, the average diesel particulate emission factor should be increased by 16.1 percent to be on an equivalent basis as the lead factor. With the incorpora- tion of the 16.1 percent increase, the ratios of E(D)/E(Pb) are calculated to be 1.8 (low estimate) and 2.8 (high estimate). B. Lead and Diesel Particulate Dispersion Characteristics Automotive lead and diesel particulate emissions have similar properties that would imply that their dispersion would be very similar. These properties are: a) both are emitted in particulate form, b) both are emitted from ground level, and c) both are emitted from vehicles of similar urban driving patterns. However, a major difference between lead and diesel particulate is their relative size. This section will discuss this difference and how it affects the relative dispersion of the two types of particulate. Diesel particulate is extremely small with well over 90 percent by mass being fine (less than 2.5 micrometers in dia- meter). _8_/9_/ This is small enough for all of the particulate to be considered suspendable.JjV Lead-containing particulate (lead salts such as PbClBr), on the other hand, is much larger, only 43 percent by mass being smaller than 9 micrometers in diameter ._10/ This same study examined the particle size distributions of both ambient and exhaust lead-containing particulate and concluded that only the 43 percent smaller than 9 micrometers was being suspended and the rest was settling out rather quickly after emission._10_/ It appeared somewhat simplistic to assume that the cutoff for suspension would be so sharp. However, an examination of the size distributions of both the ambient and exhaust particulate samples revealed that the fraction of the mass smaller than 0.6 micrometer was approximately 2.5 times larger for the ambient sample than the exhaust sample. If it is assumed that all of the particles less than 0.6 micro- meters in diameter were suspended, this would imply that about 40 percent of the lead-containing particulates was suspended. As this confirms the 43 percent figure cited above, 43 percent will be used as the percentage of lead-containing particulate that is suspended. Given that all the other source characterisitics are the same for lead and diesel particulate, the only differences between the dispersion of the two pollutants is that 57 percent of the lead ------- -15- Table 6 Fraction of Urban VMT by Mobile Source Category in 1974 Fraction of Urban VMT i 1 2 3 4 5 6 Classification LDV-G LDV-D LDT-G LDT-D HDV-G HDV-D (1974) 2/ 0 0 0 0 0 0 .826 .004 .107 .001 .036 .026 Pbi 0 0 0 0 0 0 (gr/mi) .11 .11 .27 G = Gasoline. D = Diesel. ------- -16- particulate does not stay aloft long enough to be measured by an ambient air monitor. Thus, the ratio f(D)/f(Pb) is 1.0/0.43 or 2.32. C. Ambient Lead Levels The third and last parameter to be determined before equation (3) can be used to predict ambient levels of diesel particulate is the ambient level of lead (C(Pb)). Unlike the other two para- meters, this parameter varies from city to city. For precisely that reason, this approach is able to yield diesel impacts in many cities while requiring a minimum of effort. There are two primary sources of ambient lead data available; that obtained by the National Air Surveillance Network (NASN) and that contained in the National Aerometric Data Bank (NADB).JJ_/ The NASN data will be used here because it has the greater likelihood of being representative of large-scale urban areas and large exposed populations. Many of the lead monitors submitting data to NADB are special purpose monitors located near large sources of lead emissions and would only be representative of locales near those sources. The NASN ambient lead data is shown in Table 7-A through 7-E for cities divided into five population categories. Data from a few cities known to have large stationary sources of lead emission have been omitted. In general, the lead measured at sites shown in Tables 7-A through 7-E should be nearly all due to motor vehicle exhaust emissions. In 1975, motor vehicles accounted for 89 percent of the 142,000 metric tons emitted nationwide.^/ In addition, much of the 11 percent due to stationary source emissions is concentrated in those areas which have been avoided by this study. However, to be conservative, it will be assumed that only 89 percent of the ambient lead concentrations shown in Table 7-A through 7-E are due to motor vehicle emissions. Thus, these values will be multiplied by 0.89 before being used in equation (3). D. Calculation of Ambient Levels of Diesel Particulate The necessary data is now available to calculate ambient diesel particulate levels in 1990. Once again equation (3) is: C(D) = E(D) ' f(D) . C(Pb) (3) E(Pb) f(Pb) The ratio E(D)/E(Pb) is 1.8 for the "low estimate" case, and 2.8 for the "high estimate" case. The ratio f(D)/f(Fb) is equal to 2.32 (1/0.43). C(Pb) for each city is equal to the value shown in Table 7 (A through E) multiplied by 0.89. Using these figures, ------- Table 7-A Ambinet Lead Levels in Cities with a Population Greater Than 1,000,000 ll/ City Chicago Detroit Houston Los Angeles New York Philadelphia AQCR # 67 67 123 216 24 43 45 45 Site Number 141220001 141220002 231180001 452560001 054180001 (Old #) 334680001 (New #) 334680014 397140002 397140004 Address 320 N. Clark St. 445 S. Plymouth Ct. Public Library 810 Bagby St. 434 S. Pedro St. 170 E. 121st St. 2031 Race St. 1501 E. Lycoming Ave . Station Type Center City- Commercial Center City- Commercial Suburban- Commercial Center City- Commerical Center City- Commercial Center City- Commercial Center City Residential Suburban- Residential Elev. Above Ground ( f t . ) 10 10 9 50 100 75 15 17 Lead Concentration (ug/m^)* 1.42 3.01 0.99 2.09 2.68 1.05 1.34 1.23 Annual Mean. ------- Table 7-B Ambient Lead Levels in Cities With a Population Between 500,000 and 1,000,000 ll/ City Boston Dallas Denver Kansas City, New Orleans Phoenix Pittsburg San Diego St. Louis AQCR # 119 215 36 36 MO 94 106 15 197 29 70 72 Site Number 220240001 451310002 060580001 060580002 262380002 192020002 030600002 397260001 056800001 264280001 264280002 Address JFK Bldg., Cambridge St. 2100 Young St. 414 14th St. 2105 Broadway Not available 421 Loyola Ave . 1845 F. Roosevelt County Office Bldg. Not available 1720 Market St. 215 S. 12th Blvd. Station Type Center City- Commerical Center City- Commercial Center City- Commercial Center City- Commerical Center City- Commerical Center City- Commercial Center City- Commercial Center City- Commercial Center City- Commercial Elev. Above Ground ( f t . ) 85 12 43 9 72 30 160 49 10 Lead Concentrations (ug/m3)* 0.92 3.03 0.95 1.59 0.80 1.06 2.10 0.85 1.13 1.18 1.58 00 I Annual Mean. ------- Table 7-C Ambient Lead Levels in Cities With a Population Between 250,000 and 500,000 City Atlanta Birmingham, AL Cincinatti Jersey City Louisville Oklahoma City Portland Sacramento Tucson Yonkers, NY AQCR # 56 4 79 43 78 184 184 193 28 15 43 Site Number 110200001 010380003 361220001 (A01) 312320001 182380002 372200015 372200029 381460001 056580001 030860001 337620001 (A01) Address 99 Butler St. SE Not Available Public Library, Vine St. Med. Ctr. Garage 2500 S. 3rd St. 428 W. California Not available State Office Bldg. 2221 Stokton Blvd. 24D & Palm 87 Hepperman Ave . Station Type Center City- Commericial Center City- Commercial Suburban- Industrial Center City- Industrial Center City- Commerical Center City- Commerical Center City- Commerical Center City- Commerical Center City- Commerical Elev. Above Ground ( f t . ) 20 550 45 80 15 170 11 47 100 Lead Concentrations (ug/m3) * 1.05 1.22 0.81 1.03 0.96 •8. H> VD 1.66 ' 1.02 0.81 1.05 0.75 1.16 Annual Mean. ------- Table 7-D Ambient Lead Levels in Cities With a Population Between 100,000 and 250,000 City AQCR # Baton Rouge 106 Jackson, MS 5 Kansas City, KA 94 94 Mobile, AL 5 New Haven, CT 42 Salt Lake City 220 Spokane, WA 62 Torrance, CA 24 Trenton, NJ 45 Waterbury, CT 42 Site Number 190280001 251260002 171800002 171800012 012380001 070700001 460920001 492040001 058260001 315400001 071240001 Address 3142 Evangeline St. 424 N. State St. Miami & Baltimore EPA Lab 25- Furston Rd . O.K. Bicycle Shop 270 Orange St. 610 S. 2nd East Spokan City Hall 2300 Carson St. State House and State St. City Hall 235 Grand Ave . Station Type Center City- Commercial Center City- Industrial Center City- Industrial Center City- Commercial Center City- Commercial Center City- Commercial Center City- Commercial Center City- Residential Center City Commercial Center City- Commercial Elev. Above Ground ( f t . ) 5 12 14 19 15 72 30 84 4 40 55 Lead Concentrations (ug/m3) * 0.93 0.80 0.60 0.43 0.96 1.15 0.98 0.58 2.35 0.88 1.88 o Annual Mean. ------- Table 7-E Ambient Lead Levels in Cities With Population Under 100,000 Kiev. Above Lead Concentration City Anchorage, Bethlehem, Helena, MO Jackson Co AQCR # AL 8 PA 151 142 . , MS 5 Site Number 020040003 39078002 270720001 251280001 Address 527 E. 4th Ave. Public Safety Bldg. Cogswell Bldg. Jackson Co. Health Dept. Station Type Center City- Commercial Suburban- Commercial Center City- Residential Rural- Near Urban Ground (ft.) 28 41 29 4 (ug/m3) * 1.00 0.57 0.29 0.47 I NJ Annual Mean. ------- -22- ambient diesel particulate levels in 1990 can be calculated and are shown in Tables 8-A through 8-E. Now that all the previous studies have been normalized to the same scenario and the GM work has been extended to more cities, the last step of this analysis will be to compare the results of the studies and determine what is the best estimate available of the future ambient impact of diesel particu- late emissions. V. Comparison of Results The normalized results of the various studies are contained in Tables 2,4,6, and 8-A through 8-E. Because the projections con- tained in Table 8 (A-E) include most of the cities examined by the other studies, the Table 8 data will be used as a common ruler, against which the results of the other studies will be compared. The comparison will begin with the GM work, as it will be the simplest comparison and will be followed by the PEDCo three city study and the PEDCo-Kansas City study. A. GM Study This comparison is the simplest because the methodology used in Section C is most near that of GM. The city examined both by GM and in Section IV is Chicago. Under the low diesel estimate scenario, GM would have projected a level of 10.9 micrograms per cubic meter while Table 8-A shows 11.2 micrograms per cubic meter (for the monitor examined by GM). As can be seen, these projec- tions are less than 3 percent apart. A brief conversation with GM revealed a few sources of the difference, some compensating others.12/ One, GM used a higher average lead emission factor, 0.13 g/mi versus the 0.11 g/mi factor determined in Section IV-A. Two, GM did not try to take into account sources of lead other than motor vehicles. Thus, the 0.89 factor was not used. Three, their original ambient lead level was slightly higher (3.2 versus 3.0 micrograms per cubic meter) as they used 1970 data_13_/ rather than 1975 data.7/ GM's higher lead emission factor may be due to the earlier date examined, 1970. The lead content in fuel was decreasing in that time frame .Tj The decision to take stationary source lead emissions into account is really a decision to be conservative or liberal and in this study the choice has generally been to be conservative, if possible. In general, then, the methodologies used by GM and in Section IV seem to be nearly identical and the GM results tend to confirm the results of Section IV. As the latter examines many more cities than the GM work, the results in Table 8 should be sufficient for future studies. B. PEDCo-New York, Chicago, and Los Angeles Study The basis of this PEDCo study, like that of Section IV, is the ------- Table 8-A Ambient Diesel Particulate Levels* for Urban Areas With A Population Above 1,000,000 City Chicago Detroit Houston Los Angeles New York Philadelphia AQCR # 67 67 123 216 24 43 45 45 Ambient Lead, 1975 (ug/m3)(All Sources) 1.42 3.01 0.99 2.09 2.68 1.05 1.34 1.23 Ambient Paticulate, 1990 (ug/m3) "Low Estimate" 5.29 11.22 3.70 7.76 9.96 3.90 4.99 4.56 Ambient Particulate, 1990 (ug/m3) "High Estimate" 8.36 17.73 5.83 12.32 15.80 6.19 i S3 w 7.90 ' 7.25 Annual Mean. ------- Table 8-B Ambient Diesel Particulate Levels* for Urban Areas With A Population From 500,000 to 1,000,000 Ambient Lead, 1975 City AQCR # (ug/m3)(All Sources) Boston Dallas Denver Kansas City, MO New Orleans Phoenix Pittsburgh San Diego St. Louis 119 215 36 36 94 106 15 197 29 70 70 0.92 3.03 0.95 1.59 0.80 1.06 2.10 0.85 1.13 1.18 1.58 Ambient Paticulate, 1990 (ug/m3) "Low Estimate" 3.41 11.27 3.53 5.91 2.61 3.92 7.81 3.15 4.21 4.38 5.88 Ambient Particulate, 1990 (ug/m3) "High Estimate" 5.42 17.86 5.60 9.37 4.13 6.24 i N * 12.38 ' 5.01 6.67 6.95 9.32 Annual Mean. ------- Table 8-C Ambient Diesel Particulate Levels* for Urban Areas With A Population From 250,000 to 500,000 City Atlanta Birmingham, AL Cincinnati Jersey City Louisville Oklahoma City Portland Sacramento Tucson Yonkers, NY Ambient Lead, 1975 AQCR # (ug/m3)(All Sources) 56 4 79 43 78 184 184 193 28 15 43 1.05 1.22 0.81 1.03 0.96 1.66 1.02 0.81 1.05 0.75 1.16 Ambient Paticulate, 1990 (ug/m3) "Low Estimate" 3.90 4.54 3.02 3.83 3.57 6.16 3.78 3.02 3.90 2.80 4.31 Ambient Particulate, 1990 (ug/m3) "High Estimate" 6.18 7.19 4.77 6.07 5.65 9.78 & 6.00 ' 4.77 6.19 4.42 6.83 Annual Mean. ------- Table 8-D Ambient Diesel Particulate Levels* for Urban Areas With A Population From 100,000 to 250,000 Ambient Lead, 1975 City AQCR # (ug/m3)(All Sources) Baton Rouge Jackson MS Kansas City, KA Mobile, AL New Haven, CT Salt Lake City Spokane , WA Torrance , CA Trenton, NJ Waterbury, CT 106 5 94 94 5 42 220 62 24 45 42 0.93 0.80 0.60 0.43 0.96 1.15 0.98 0.58 2.35 0.88 1.88 Ambient Paticulate, 1990 (ug/m3) "Low Estimate" 3.46 2.97 2.24 1.60 3.57 4.28 3.65 2.15 8.74 3.27 6.70 Ambient Particulate, 1990 (ug/m3) "High Estimate" 5.48 4.71 3.46 2.54 5.65 6.78 5.78 3.42 13.85 5.19 11.08 I NJ Annual Mean. ------- Table 8-E Ambient Diesel Particulate Levels* for Urban Areas With A Population Under 100,000 Ambient Lead, 1975 Ambient Paticulate, 1990 City AQCR # (ug/m3)(All Sources) (ug/m3) "Low Estimate" Anchorage, AK 8 1.00 3.65 Bethlehem, PA 151 0.57 2.07 Helena, MO 142 0.29 1.06 Jackson 5 0.47 1.67 County, MS Ambient Particulate, 1990 (ug/m3) "High Estimate" 4.44 2.52 1.29 2.76 i N3 Annual Mean. ------- -28- use of lead as a surrogate. As such, one would expect the results of the two studies to be similar. However, this is not the case. The PEDCo results for New York are much higher (86-154 percent) than those determined in Section IV as are PEDCo"s Chicago results to a lesser extent. A possible reason for this might be the use of different monitors within each city. However, an examination of the actual sites modelled (Tables 6 and 8-A) reveals that two sites, one in New York and one in Los Angeles, were modelled in both studies. In both cases, the PEDCo results were higher, 154 percent (New York) and 18 percent (Los Angeles). The latter error is not large given the type of projections being made here, but the former was too large to ignore and PEDCo's methodology was examined to identify possible sources of the differences. One primary difference was found between PEDCo's metodology and that used in Section IV. In Section IV, ambient lead levels are modified by two factors, one related to emissions and one related to dispersion characteristics. PEDCo used two similar factors. However, the base ambient lead level was not measured, but calculated from ambient levels of total suspended particulate (TSP). A different constant fraction of TSP levels was assumed to be lead (or lead salts) in each of the three cities, based on referenced studies in New York and Los Angeles. Both of these references were examined. The Los Angeles study concluded that 13 percent of ambient TSP levels in Los Angeles were automotive-related.14/ It appeared from the wording of the report that this 13 percent only included leaded exhaust particulate and not tire particulate or reentrained road dust.JL4/ This was also PEDCo's interpretation judging from their use of the 13 percent figure ,_5_/ The New York study, on the other hand, concluded that 20-25 percent of New York's TSP levels were automotive-related._15y A statistical method was used to correlate ambient TSP levels with ambient lead levels. Rather than express lead levels as a fraction of TSP levels, however, the report's results were actually in terms of an 'x' microgram per cubic meter increase in ambient elemental lead levels coincides with a 'y' microgram per cubic meter increase in TSP levels.15/ This type of analysis would definitely include reentrained dust and other-than-exhaust automotive particulate. However, PEDCo interpretted the report's conclusion to only refer to exhaust (lead-salt) particulate and assumed that 21 percent of ambient TSP levels were lead-containing particulate. This would appear to be the source of the difference between the PEDCo New York results and that of Section IV. This error can be remedied by determining the actual percen- tage of TSP levels due to "exhaust" particulate. A reexamination of the original New York study revealed that about 8 percent of New York TSP levels consisted of lead salts from motor vehicles.15/ As ------- -29- PEDCo used 21 percent, they overestimated the actual figure by a factor of 2.66. PEDCo's New York results can simply be divided by 2.66 to remove the error and this has been done in Table 9. As can be seen, the modified PEDCo result for the E. 121st St. site is now 6.0 microgram per cubic meter, which compares very well with the Table 8-A result of 6.19 microgram per cubic meter. Thus, the two studies now compare very well in New York and moderately well in Los Angeles. However, the PEDCo results for Chicago were partly based on the erroneous 21 percent figure used for New York. As no similar study was available for Chicago, PEDCo assumed that Chicago's motor vehicle contribution to TSP levels would be halfway between that of New York and Los Angeles, or 17 percent. Given that the New York percentage is now 8, the same assumption would yield 11 percent for Chicago or a reduction of 35 percent. Thus, the PEDCo results for Chicago should be multiplied by 0.65 to adjust for this error. The adjusted Chicago results are also shown in Table 9. As can be seen by comparing the Chicago results in Tables 8-A and 9, four out of five of the PEDCo monitors fall within the range of the two Table 8-A monitors and agreement can be said to be quite good. Because this PEDCo study examined a number of monitors in each city, further analysis of the locations of these monitors was performed to determine any possible localized effects due to heavy traffic nearby. Since the modified PEDCo results agree very well with the Section IV results, any conclusion made concerning the PEDCo sites should also apply to the sites modelled in Section IV. Numerous calls were made to state, local, and EPA regional offices to determine the location of the SAROAD sites modelled by PEDCo. The results are shown in Table 10. EPA guidelines for TSP monitors were published recently and contained minimum distances that a monitor should be located from a road to be representative of large-scale impact s .J_6_/ These minimum distances are 1) 15 meters above and 5 meters away from the road, 2) 2 meters above and 25 meters away from the road, or 3) any point lying on a straight line between these two positions. As can be seen for New York, all five monitors lie well outside these limits. Also, while the traffic counts of the nearest streets are significant, none can be termed 'heavily—travelled' by New York standards. Thus, the projected diesel particulate levels at these sites should be very represenative of large-scale areas and not representative so-called localized impacts. The Los Angeles sites are generally closer to the road then the New York sites. Two out of four sites for which locations are available meet the EPA guidlines. Again as in New York, none of the nearest roads are exceptionally busy. From an examination of the ambient diesel impacts at these five sites, one finds that ------- -30- Table 9 Projected Ambient Diesel Particulate Concentrations in 1990 Revised PEDCo Results City New York New York New York New York New York Torrance, Los Angeles Long Beach, Los Angeles Los Angeles Pasadena Pasadena Chicago Chicago Chicago Chicago Chicago Site Address Steimnan Hall ,W. 141 St and Convent Ave . 170 E. 121 St. Central Park Arsonal, 5th Ave., and 64th St. 240 2nd Ave . Pier 42, Morton St. and Hudson River 2330 Carson St. 2655 Pine Ave. 434 S. Pedro 1196 East Walnut Kech Laboratories, Cal. Inst. of Tech. 3500 E. 114 St. 1947 W. Polk 9800 S. Torrence Ave. 538 S. Clark St. 4015 N. Ashland Ave. EPA Scenario Low 3.7 3.9 3.6 4.6 4.4 9.6 10.0 12.0 10.1 11.0 17.3 8.5 9.3 8.8 7.3 (ug/m3) High 5.7 6.0 5.5 7.1 7.0 14.9 15.4 18.7 15.6 17.1 26.9 13.2 14.4 13.7 11.3 ------- Table 10 Saroad Monitoring Sites City New York New York New York New York New York Torrance, Los Angeles Long Beach, Los Angeles Los Angeles, Los Angeles Pasadena, Los Angeles Site Address SAROAD Code Steinman Hall 334680057F01 W. 141 St. , and Convent Ave . 170 E. 121 St. 334680014P01 Central Park 334680005H01 Arsonal, 5th Ave., and 64th St. 240 2nd Ave. 334680010H01 Pier 42 ? Morton St. , and Hudson River 2330 Carson St. 058260001P01 2655 Pine Ave. 05410001F01 434 S. Pedro 054180001101 1196 East Walnut 055760004101 Elevation Above Ground 22.9 m (75 ft.) 22.9 m (75 ft.) 13.73 m (45 ft.) 18.3 m (60 ft.) 7.63 m (25 ft.) 1.22 m (4 ft.) 7.63 m (25 ft.) 27.4 m (89.8 ft.) 5.5 m (18 ft.) Distance From Large Road 91.5 m (300 ft.) 30.5 m (100 ft.) 30.5 m (100 ft.) 15.25 m (300 ft.) 91.5 m (300 ft.) Not Available 1.83 m (6 ft.) 5.0 m (16.4 ft.) 17 m (55.7 ft.) Vehicle Count (Vehicle/day) Comments 12,100 16,500 17,900 26,600 Air Resource Board lists 27.45m (90 ft.) above ground, i i — ' 16,800 City lists 4.58m (15 ft.) above ground . 15,000 15,000 13,500 18,000 ------- Table 10 (con't) PEDCo TSP Monitoring Sites City Lennox , Los Angeles Chicago Chicago Chicago Chicago Chicago Site Address 11408 Blvd. 3500 1947 9800 Ave . 538 S 4015 Ave . La Cienega E. 114 St. E. Polk S. Torrence . Clark St. N. Ashland SAROAD Code 05390000101 14122002H01 141220033F01 141220005H01 141220005H01 141220004H01 Elevation Above Ground 7.0 m (23 ft.) 9.46 m (31 ft.) 4.57 m (15 ft.) 4.88 m (16 ft.) 39.9 m (133 ft.) 19.2 m (64 ft.) Distance From Large Road 19 m (62.3 ft.) 24.4 m (80 ft.) 30.5 m (100 ft.) 21.35 m (70 ft.) 9.15 m (30 ft.) 3.6 m (12 ft.) Vehicle Count (Vehicle/day) 25,000 Not Available 4,674 9,400 11,600 25,100 Comments i u> KJ 1 ------- -33- the largest impact is at the San Pedro site, which is the furthest from the least-travelled road. From this observation, it would be difficult to argue that the other sites were overly influenced by heavy traffic. The result for Los Angeles, then, is the same as that for New York, the monitors appear to be very representative of large-scale impacts. The Chicago monitors are slightly more difficult to analyze. Four out of five are within the EPA guidelines, though two of these four monitors (Polk and Torrence) are quite near lightly- travelled streets. The Clark St. monitor, on the other hand, is well away from a lightly-travelled street. The impact at this monitor (Table 9) is no different than that at the three monitors which are nearer the road. Only the 114th Street monitor has an usually high impact associated with it. This monitor is actually quite far from the street, though the exact traffic count of the street is not known.As PEDCo' s methodology was based on TSP levels, the TSP level at this site was twice that of the others in Chicago and upon investigation was found to be 165 micrograms per cubic meter.5/ It cannot be determined whether the motor vehicle contribution at this monitor is also twice that at the other four monitors. However, as the resulting diesel particulate level is almost 50 percent higher than that at any other monitor in any city, it appears likely that something unusual is occurring at that monitor. The use of the 114th Street projection should therefore be used with caution. Otherwise, the results of the other four monitors appear to be free from local impacts like heavy traffic nearby and should be representative of large-scale impacts. Given that the great majority of the monitors examined by PEDCo appear to be representative of large-scale impacts and the modified PEDCo results closely match those of Section IV, it would seem reasonable to assume that the great majority monitors examined in Section IV are also representative of large-scale impacts. As such, either set of projections could be used to project ambient diesel particulate impacts over large-scale urban areas. C. PEDCo-Kansas City Study The final comparison to be performed is between the PEDCo- Kansas City results (Table 2) and the projection for Kansas City developed in Section IV (Table 8-B). As can be seen, the PEDCo projections are less than half those in Table 8-B. Somewhat aggrevating this difference is the fact that the Section IV pro- jection for Kansas City, MO., is the lowest to be found in Tables 8-A, 8-B, and 8-C. In other words, with respect to the impacts in other large cities, the Section IV Kansas City impact appears to be too low rather than too high. Given this and the fact that the impacts shown in Table 8 are consistent with PEDCo's three city study and GM's work, it would appear that PEDCo underestimated the diesel's impact in Kansas ------- -34- City. This criticism in fact, has already been made in a previous comparison of carbon monoxide and diesel particulate emissions in Kansas City.llJ The problem most likely lies with the use of AQDM, which would be expected to underestimate line sources, such as motor vehicle .J_8/ Because of the discrepancy between PEDCo ' s Kansas City results and that of the other studies, the impacts shown in Table 2 should not be used as valid projections of future diesel impacts. VI. Conclusions 1. In the PEDCo-Kansas City study, the use of AQDM had probably underestimated future ambient particulate levels from diesel emission in Kansas City. 2. In the PEDCo three-city study, an error was made that overestimated the impacts in New York and Chicago. After correc- tion of this error, however, the resulting impacts appear to be reasonable for further use. In addition, the locations of the monitors modelled were such that the projections should be repre- sentative of large-scale urban impacts. 3. The GM work has been repeated, essentially confirmed, and extended to many other cities in the U.S. Along with the PEDCo three-city results, these Section IV projections appear to be among the best available and could be used in further studies. ------- -35- References \J "Draft Regulatory Analysis, Light-Duty Diesel Particulate Regulations," OANR, OMSAPC, EPA, Dec. 22, 1978. _2_/ "Air Quality Assessment of Particulate Emissions from Diesel Powered Vehicles," PEDCo Environmental Inc. for EPA, March 1978, Contract No. 68-02-2515. _3_/ "Summary and Analysis of Comments on the Notice of Proposed Rulemaking for Light-Duty Diesel Particulate Regulations for 1981 and Later Model Year Vehicles," EPA, OANR, OMSAPC, ECTD, SDSB, October 1979. kl Personal communications with Jim Throgmorton, PEDCo, June 15, ~ 1979. _5_/ "The Impact of Future Diesel Emission on the Air Quality of Large Cities," FED Co Environmental for the EPA, February 1979, Contract No. 68-02-2585. 6J "General Motors Response to EPA Notice of Proposed Rulemaking on Particulate Regulation for Light-Duty Diesel Vehicles," Attachment 4, General Motors, April 19, 1979. ]_/ "Environmental Impact Statement for Lead," OANR, OAQPS, EPA, September 1978. 8/ Grolicki, P.J., and C.R. Begeman, "Particle Size Variation in ~~ Diesel Car Exhaust," SAE 790421. 9/ Schreck, Richard J. et al, "Characterization of Diesel Exhaust Particulate Under Different Engine Load Conditions," Presented at 71st Annual Meeting of APCA, June 25-30, 1978. 10/ Huntzicker, James J. et al, "Material Balance for Automobile Emitted Lead in the Los Angeles Basin," Environmental Science and Technology, Vol. 9, 1975. ll/ "Standard Support and Environmental Impact Statement: Na- tional Ambient Air Quality Standard for Lead-Emissions, Air Quality, and Environmental Impact," Appendices A through Y, Mitre Corporation, MTR-7525, Vol. II. 12/ Telephone conversation with Richard Klimisch of General Motors, July 10, 1979. 13/ "Air Quality Criteria for Lead," EPA, 1977, EPA-600/8-77-017. ------- -36- Reference (cont'd.) 14/ Hidy, G. M. and S. K. Friedlander, "The Nature of the Los Angeles Aerosol," Proceedings of the Second International Clean Air Congress, 1971. 15/ Kleinman, Michael T., "The Apportionment of Sources of Air- borne Particulate Matter," Doctoral Dissertation at New York University, New York, N.Y., June 1977. 16/ "Air Quality Surveillance and Data Reporting," 43 FR 34892, August 7, 1978. 17/ Rykowski, Richard A., "Relative Impact of CO and Particulate on Air Quality," EPA Memorandum to Charles L. Gray, Jr., Director, ECTD, August 1979. 18/ Neilgan, Robert E., Director, Monitoring and Data Analysis Division, OAQPS, OANR, "Information Concerning Particulate Emissions from Nonmobile Sources," EPA Memorandum to Charles L. Gray, Jr., Director, Emission Control Technology Division, OMSAPC, OANR, July 11, 1979. ------- |