Remote Sensing Study of Emissions from Motor Vehicles in the

Remote sensing was employed for the first time to measure nitric oxide (NO) levels of on-road light-duty motor vehicles of the Metropolitan Area of Me...
0 downloads 0 Views 175KB Size
Download PDF
Environ. Sci. Technol. 2003, 37, 395-401

Remote Sensing Study of Emissions from Motor Vehicles in the Metropolitan Area of Mexico City I . S C H I F T E R , * L . D IÄ A Z , J . D U R AÄ N , E . G U Z M AÄ N , O . C H AÄ V E Z , A N D E. LO Ä PEZ-SALINAS Instituto Mexicano del Petro´leo, Competencia de Estudios Ambientales, Eje Central 152, Me´xico, D.F., 07730 Mexico

Remote sensing was employed for the first time to measure nitric oxide (NO) levels of on-road light-duty motor vehicles of the Metropolitan Area of Mexico City (MAMC). The sensor placed at 12 different sites also measured the concentration of CO2, CO, and total hydrocarbons (THC) in the exhaust emissions. A database was compiled containing 122 800 readings, of which 84 650 (69%) records were valid emissions measurements. CO, HC, and NO valid readings were 68.9, 63.4, and 62.9%, respectively, of the total attempted measurements. Furthermore, 42 822 vehicles were number-plate-matched to model year with the information provided by the Inspection/Maintenance Program. The mean emissions of total valid readings for CO, HC, and NO were determined to be 1.31 vol %, 440 ppm (propane), and 914 ppm, respectively. In 1991 and 1994, remote sensing measurements of CO and HC tailpipe emissions were performed in the MAMC in five different locations (30 000 valid readings). Large drops in both pollutants were observed for the intervening years, but sufficient vehicle information was not available at that time to fully explain the observed trends. Compared with those reports, our results point out to a steady decrease in CO and HC exhaust emissions with vehicle model year. The fleet emissions measured exhibit a γ-distribution, with 10% of the most polluting fleet studied being responsible for 45%, 25%, and 29% of the CO, HC, and NO emissions, respectively. NO emissions in taxis are the highest among the vintage of vehicles, a matter of concern since according to the distance traveled per year, they represent 22% of the total activity in the MAMC.

Introduction Mexico City and its metropolitan area form one of the largest urban concentrations in the world with a population of over 18 million inhabitants and a fleet of gasoline vehicles that surpasses 3 million units (25% at the total national population). The area sits in a basin 2240 m above sea level surrounded by mountains that rise 1 km or more above the basin. High elevation, low pollutants dispersion rates, and intense sunlight are important factors in ozone formation. In a typical meteorological pattern, cold air drains off the mountain slopes, resulting in nearly stagnant air over the city. During the 1980s, the environmental authorities realized * Corresponding author e-mail: [email protected]; (+52)3003-8507; fax: (+52)3003-8484. 10.1021/es0207807 CCC: $25.00 Published on Web 12/06/2002

 2003 American Chemical Society

phone:

that ozone would become a major problem and that vehicular emissions control would be required to maintain a satisfactory air quality. The Vehicle Emissions Verification Program in the MAMC, launched in 1989, mandated inspection and maintenance (I/M) twice a year to ensure that vehicles operated as clean as their engine and emission control systems are designed (1). In 1993, the I/M was upgraded and centralized: the equipment and test procedures used met the BAR90 (California Air Resources Board Bureau of Automotive Repair) standard (2). The current dynamic I/M test proceeds in two stages. In the first one at 24 km/h, load is applied to the vehicle that equates 50% of the maximum acceleration rate that the vehicle has in the U.S. Federal Test Procedure, FTP75 (3). In the second one, as in the case of the ASM5015 test procedure, the vehicle is accelerated to 40 km/h for 60 s. During the first semester of 1989, the MAMC initiated a “One-day-without-a-car” program. Data available at that time showed the highest ozone levels to occur between October and March each year. To overcome this situation, the circulation of all the vehicles during the winter months was restricted by 1 day a week, and later the program became year-round. During an emissions contingency, only the least polluting vehicles are allowed to circulate. From May 1998, contingency is applied in the MAMC when O3 concentration is higher than 0.282 ppm (240 points from IMECA index) (4). Hence, different certificates are issued to the vehicles, through a mixture of emissions and technology level, each with a different and highly visible windshield sticker. In 1999, the I/M Program mandated a maximum HC limit of 300 ppm for all light-duty gasoline vehicles model year (MY) 1990 and before, with a limit of 200 ppm for 1991-1992 vehicles, and 100 ppm for the 1993 up to present day MY. The CO limit was set at 3.0 vol % for vehicles up to the 1990 MY, 2.0% for 1991-1992 models, and 1% for 1993 and later cars. In January 2000, for the first time in the MAMC, limits of 2500 ppm were fixed for NOx emissions up to 1990 MY, 1500 ppm for 19911992 MY, and 1200 for 1993 and later models. Moreover, 1993-1995 MYs were obliged to replace their used threeway catalyst (TWC) to meet the 2000 I/M exhausts limits. In 2000, the MAMC environmental protection agencies estimated that, for 1998, Mexico on-road vehicles contributed 98, 77, and 35% of CO, NOx, and volatile organic compound (VOC) emissions, respectively. Moreover, 82% of the year, the air quality standards for ozone, a major component of urban smog, produced by the photochemical reaction of NOx and HC were exceeded (5). This condition not only adds to the overall ambient air quality problem but also produces considerably higher concentrations of pollutants in the urban environment, affecting virtually all the inhabitants of the MAMC. It is important, therefore, to collect data in order to evaluate the effectiveness of the different emission control programs, and for that, decision-makers need abundant data to characterize populations. A large number of samples are necessary to make meaningful comparisons between variables such as vehicle age and technology, environmental conditions, vehicle use, I/M programs, and fuel formulation. One way to measure the emissions of a large numbers of vehicles is to use remote sensing (RSD) of on-road vehicles (6). Remote sensing has been used for over a decade as an effective means of identifying high-polluting vehicles, especially in terms of CO and total hydrocarbon (THC) measurements (6, 7). More recently, RSD was employed to measure NH3 levels in the exhaust of Los Angeles vehicles (8). Large databases containing remote sensing emissions VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

395

data on hundreds of thousands of vehicles have been compiled in a time- and cost-efficient manner (9, 10). These data can be useful for computing emission inventories. The RSD technology is a useful tool for monitoring the composition of vehicle exhaust plumes as the vehicle passes a roadside testing station under real driving conditions. A similar device was used in 1991 and 1994 by the University of Denver to monitor on-road emissions of CO and HC (as propane) at five sites within the MAMC (11, 12). Furthermore, in 1995, the Metropolitan Area of Monterrey, located in the northeast corner of Mexico, was also monitored. Data obtained in Monterrey showed similar CO and HC emissions to those observed in Mexico City in 1994 but were 2-4 times greater than those recorded in Baltimore, MD, in the same year (12). In 1999, RSD measurements were obtained at the United States-Mexico border (13). It was found that CO, HC, and NO readings at both the Tijuana-San Diego crossings (Otey Mesa and San Ysidro) were lower than the 1996-1997 readings at the same crossing (14). In this work, we have collected RSD measurements from 12 different sites in the MAMC; two of the places were the same monitored in 1991 by the University of Denver. Measurements included for the first time NO emissions, not performed in the former studies. The objective of this study is to assess the evolution of on-road polluting emissions with variations in vehicle MY distributions.

Experimental Section Measurements of motor vehicle exhaust emissions were carried out between June 5 and September 15, 2000. The data were collected from Monday to Friday, between 8:00 and 16:00 h. In the present work, we employed an AccuScan remote sensing system from Environmental Systems Products, CT. The instrument is made up of a nondispersive infrared component for CO2, CO, and HC detection and a dispersive ultraviolet spectrometer for NO measurement (6, 7). The mobile unit included the equipment required to measure speed and acceleration as well as license plate recognition. The primary combustion gases (HC, CO, and CO2) are measured simultaneously along the same optic path to ensure the proper application of the combustion gas equations. To avoid interference between vehicles, the unit was capable of completing the vehicle emission measurement within 0.6 s and all other measurements for a given vehicle including emissions, speed, acceleration, and plate image within 1 s. The RSD is designed to generate and monitor a nondispersive infrared and ultraviolet beam emitted and reflected approximately 10-18 in. above ground, preferably across a single lane road. Gasoline, diesel, or other fossil fuel powered vehicles drive through this beam, and the exhaust interferes with this transmission of the beam. Quantifying the interference enables the calculation of tailpipe concentrations of CO, HC, CO2, and NO. A camera simultaneously captures a digitized video image of the rear of the vehicle and its license plate. The RSD unit takes multiple rapid readings for each vehicle to characterize the exhaust plume profile and evaluate whether a valid measurement of a vehicle’s exhaust has been achieved. The criteria included how much vehicle exhaust plume was available for the duration of a 0.6-s sampling period, evaluation of whether plume measurements were consistent with normal plume dissipation, and correction for changes in background concentrations of emissions. The measuring instruments were calibrated daily with a certified mixture of gases containing CO, CO2, propane, and NO with the balance being nitrogen. HC measurements were expressed in their “n-hexane ppm equivalents”, except where indicated (e.g., n-propane equivalents). The CO tolerance was 10% or 0.25% (whichever is greater) for all expected concentrations below 3.0% and 15% for all CO expected 396

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

FIGURE 1. Location of RSD sites within the MAMC. concentrations above 3.0%. In the case of HC, the tolerance was 150 ppm or 15% of the expected HC concentration (whichever was greater) throughout the range of HC concentrations. The NO tolerance was 250 ppm or 15% of the expected NO concentration (whichever is greater) throughout the range of NO concentrations. The unit was equipped with a speed and acceleration measurement system that uses low-energy lasers to calculate the speed to within (0.8 km/h and acceleration to within (0.5 km h-1 s-1 at the moment exhaust plume is measured. The system captures emissions readings and rear pictures of vehicles that pass through the RSD infrared beam. The video and emissions readings taken are stored directly on a removable media disk and can be used for future reference. The system automatically compensates for background emissions and uses an automatic data analysis routine, which triggers and attempts a measurement each time the infrared beam is blocked. As shown in Figure 1, 12 sites were selected within the MAMC, representative of all the on-road vehicle traffic. Measurements were distributed to represent different socioeconomic wards of the MAMC according to data provided by the environmental authorities of the MAMC and characteristic of main access routes to the city. Detailed information on the site locations can be found in the Supporting Information section. All the sites were prequalified for singlelane operation with space for the RSD equipment to be deployed without disrupting traffic flow. The instrument measured exhaust emissions from vehicles typically traveling between 20 and 40 km/h. Two sites, IMP and Polanco, were the same as those monitored by the University of Denver in 1991. Following the days of data collection, the video tapes were analyzed for license plate identification. Both Mexico City and Mexico State governments provided the I/M database for the first and second semester of 1999. The information included, among other things, vehicle MY according to the license plate and the certificate obtained in the I/M test. Only number-plates in readable conditions were matched against registration records. An attempted measurement is defined as a beam block, followed by 0.5 s of data collection. If the data collection period is interrupted by another beam block from a close incoming vehicle, the measurement attempts are aborted, and an attempt is made at measuring the second vehicle. In this case, the beam block from the first vehicle is not recorded as an attempted measurement. Invalid measurement attempts arise, for example, when the vehicle plume is highly diluted, higher speed, and accelerations rates or the reported error in the ratio of the pollutant to CO2 exceeds a preset limit. Criteria to invalidate a reading were taken from ref 15.

TABLE 1. Number of Attempted and Valid Readings, Submitted and Matched Number of Plates per Sitea valid readings

submitted plates

matched plates

date

site

attempted

CO-CO2

HC

NO

CO

HC

NO

CO

HC

NO

Jun 5-9 Jun 12-16 Jun 19-23 Jun 26-30 Jul 3-7 Jul 10-14 Jul 17-21 Jul 24-28 Aug 7-11 Aug 14-18 Aug 21-25 Sep 11-15

IMP Ecatepec Polanco Nezahualcoyotl Tlalpan Tultitlan Tlalnepantla Pedregal Azcapotzalco V. Carranza Cuautitlan Huixquilucan

10 554 10 693 13 690 13 203 18 324 6 050 5 286 6 522 9 973 9 459 10 221 8 825

6 799 7 510 10 938 7 302 11 113 5 056 4 019 4 486 7 651 5 149 6 883 7 744

5 991 6 759 10 610 6 579 10 419 4 262 3 674 4 316 7 159 4 792 5 737 7 531

6 014 6 943 10 099 6 616 10 261 4 632 3 632 4 214 6 974 4 475 5 987 7 422

4 665 6 452 10 344 6 516 10 034 4 220 3 378 4 435 6 851 4 630 3 855 7 294

4 359 5 878 10 069 5 911 9 457 3 690 3 174 4 264 6 422 4 324 3 555 7 101

4 382 6 011 9 922 5 936 9 291 3 940 3 105 4 143 6 242 4 039 3 600 6 978

2 293 3 889 6 381 3 749 6 048 2 582 2 068 2 804 4 210 3 008 1 766 4 024

2 150 3 499 6 170 3 347 5 685 2 244 1 893 2 670 3 930 2 807 1 589 3 937

2 162 3 590 5 925 3 374 5 591 2 379 1 852 2 607 3 842 2 619 1 605 3 859

12 2800

84 650 69

77 829 63

77 269 63

72 674 59 86

68 204 56 88

67 589 55 87

42 822 35 51 59

39 921 33 51 59

39 405 32 51 58

total attempted (%) valid (%) submitted (%)

TABLE 2. Speed, Acceleration, and CO, HC, and NO Emissions Summarya average emissions (no. of vehicles) valid readings (84 650) submitted plates (72 654) matched plates (42 822)

a

average SD (95% CI average SD (95% CI average SD (95% CI

CO (vol %)

HC (ppm hexane)

NO (ppm)

speed (km/h)

acceleration (km h-1 s-1)

1.31 0.39 0.22 1.28 0.41 0.23 1.34 0.42 0.24

220 75.6 42.8 212 74.5 42.2 218 75.2 42.6

914 217 123 914 219 124 993 216 122

29.33 4.72 2.67 29.43 4.90 2.77 29.56 4.80 2.72

2.01 0.70 0.40 2.06 0.69 0.39 2.05 0.68 0.38

Standard deviation and confidence interval (CI) based on 12 sites average.

Results and Discussion Ideally, the vehicles measured should be traveling at a slightly increasing moderate speed on a level road or preferably a slight grade (13). CO, and to a lesser degree HC, emissions can increase dramatically under moderate to high loads encountered when vehicles accelerate at moderate high speeds. HC emissions can also increase during decelerations (16). Nitric oxide is formed primarily in the post-flame gases during combustion in the engine cylinder. The kinetics of this reaction is highly dependent on gas temperature, with elevated temperatures leading to high NO formation rates. Nitric oxide emissions are also increased with the engine operating under load and at air/fuel ratios (A/F) slightly lean of stoichiometry. Speed and acceleration distributions were recorded for the total valid measurements. An average speed of 29 km/h, slightly under the 32 km/h previously reported in the 1991, was measured. It can be noted that vehicles run an average of 2 km h-1 s-1. About 85% of the vehicles were operated in an acceleration mode between 1 and 6 km h-1 s-1, less than 7% were in a deceleration mode, and less than 7% were in cruising mode. A 0.36 km h-1 s-1 acceleration rate has been used extensively in defining driving modes (17-19). Table 1 describes the number of valid readings, submitted vehicles, and matched plates per site, while Table 2 outlines the data reduction process beginning with the number of attempted measurements and ending with the number of records containing both valid emissions measurements and vehicle registration identification. In total, 122 800 measurements were collected during the sampling periods at the 12 locations, of which 84 650 (69%) records were valid for CO and CO2 emissions measurements but only 72 269 had readable license plates. The valid attempts were reduced to

69, 63, and 63% of the total attempted measurements of CO, HC, and NO, respectively. The total number of matched plates was 42 822 records. From these data, 76% were identified as cars, 9% were taxis, 14% light-duty trucks (pickup, vans, and sport utility vehicles), and only 1% medium- and heavy-duty gasoline trucks. Until 1999 the vehicle’s owner could choose in which Federal Entity to verify polluting emissions. Because of the lack of centralized operational control of the two programs, this added flexibility caused control to be lost, and vehicles that should have been tested apparently disappeared from the program. It can be expected that the vehicles that disappeared were mainly those that would have had difficulty in passing their corresponding emissions test requirements. Average CO in this study was 1.31%, and the median was 0.33%. The distribution of CO by percent category from valid measurements and percent contribution to the total emissions, by a given category, are shown in Figure 2. The contribution to total emissions can be obtained when the number of vehicles in a given CO category is multiplied by the average emission for that range. From the sample studied, data indicate that approximately 85% of vehicles contribute to 43% of the total CO and that 10% of the high emitters contribute to 45% of total CO emissions and would be under the uppermost 3 vol % emission mandated in the I/M Program. In the results obtained by the University of Denver in 1991, an average of 4.3 vol % CO was reported with a median of 3.8% and 30% of vehicles having emissions below 2 vol % of CO (10). To compare the HC emissions with other studies, nhexane equivalents were transformed to propane equivalents. The frequency distribution of HC (propane) by category and the percentage of the contribution to total emissions by a VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

397

FIGURE 4. Fraction of vehicles and their fractional NO contribution from total valid readings. FIGURE 2. Fraction of vehicles and their fractional CO contribution from total valid readings.

TABLE 3. Average Emissions of Identified Vehicles with Model Year at Each Site

site IMP Ecatepec Polanco Nezahualcoyotl Tlalpan Tultitlan Tlalnepantla Pedregal Azcapotzalco V. Carranza Cuautitlan Huixquilucan

1.1 1.68 0.86 1.67 1.19 1.56 1.6 0.81 1.29 1.58 2.03 0.68

158 243 160 336 213 213 254 129 205 257 343 99

averagea SDa (95% CIa

1.34 0.42 0.24

218 75.2 42.6

a

FIGURE 3. Fraction of vehicles and their fractional HC contribution from total valid readings. given category are shown in Figure 3. From these data, it can be noted that 63% of the vehicles have emissions lower than 350 ppm and that 21% are in the range of 350-1050 ppm, with only 8% exceeding 1050 ppm. The average value for the entire fleet is 440 ppm HC, (propane) with a standard deviation of 151.2 and a median of 206 ppm. The 1991 data report an average of 2100 ppm for HC and a median of 1100 ppm, with 59% of vehicles with emissions exceeding 1000 ppm (11). In the present work, it was found that only 8% of the vehicles exceeded 1050 ppm but contributed to 25% of the total HC emissions. The frequency distribution of NO by category and the contribution to total emissions is shown in Figure 4. The average emission of NO is 914 ppm, standard deviation of 217, with a median of 415 ppm and 10% of the vehicle population exceeding 2500 ppm and contributing to 79% of the total NO emissions. In 1991, no evaluation of NO was performed; therefore, there are no previous RSD measurements to assess the improvements in NO control in the MAMC. RS measurements of NO emissions from on-road 398

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

HC CO (ppm NO speed acceleration (vol %) hexane) (ppm) (km/h) (km h-1 s-1) 1086 1539 701 883 954 1230 1001 955 896 945 900 821

34.2 34.7 23.1 24.9 31.4 36.9 26.5 25.1 25.1 31.5 26.8 34.5

993 29.56 216 4.80 122 2.72

model year

1.5 1.97 1.34 1.84 1.5 2.37 2.62 3.67 2.22 1.27 1.81 2.44

1993 1989.3 1994.2 1990.1 1992.3 1990.1 1991 1994.2 1992.2 1991.5 1989.4 1994.4

2.05 0.68 0.38

1991.81 1.87 1.06

Considering the averages of the 12 sites.

LDMVs show considerable variation. Zhang et al. reported mean levels of 500 ppm (standard deviation 800 ppm) obtained during a 1994 study in Denver (7). In 1996, Jime´nez et al. measured NO levels (average, 321 ppm; standard deviation, 525 ppm) in exhaust emissions from vehicles traveling at an average velocity of 13.9 m/s on a Los Angeles street (20). Popp et al. reported mean NO levels of 400 ppm with 74% of vehicles below 500 ppm during a 1997 remote sensing study in the Chicago area (21). It is known that a common practice to obtain the I/M Certificate in the MAMC has been to tune up the vehicle “Late-and-lean” with late ignition timing and lean fuel/air mixture, and the vehicle would not be properly retuned after passing the test. These techniques reduced the engine power and temporarily decrease HC and CO emissions but certainly increase NO emissions. Average emissions of the matched vehicles at each site are shown in Table 3. Variations in averages of each site show the correlation to the income level of the zone and vehicles passed through the sensor. Huixquilucan, Polanco, and Pedregal show the lowest emissions and newest MY. The MYdistribution for the matched-plate vehicles is presented in Figure 5, along with the model year distribution for the whole MAMC fleet as reported in 1998 (5). The main

FIGURE 5. Comparison of the fleet distribution by model year observed with the remote sensing and that reported in 1998 vehicle inventory.

FIGURE 6. CO average emissions as a function of vehicle model year. Error bars represent (95% confidence interval for the average of 12 sites.

FIGURE 7. HC average emissions as a function of vehicle model year. Error bars represent (95% confidence interval for the average of 12 sites.

explanation for the difference in distributions (also noted by Sjo¨din in a RSD study in Sweden) is that cars of recent model years are more frequently driven than older cars (22). In other words, the RSD sampling reflects to some extent also the average vehicles miles traveled by each model year, as opposed to the bars of the vehicle’s inventory, which represents only the number of cars for a given model year. The average age of the measured fleet was estimated to be approximately 8 yr, i.e., 1 year older than the overall fleet. We assume, therefore, that the behavior of means in valid readings, submitted vehicles, and matched plates as distribution in vehicles measured are representative of the present vehicle fleet in the MAMC. The resulting average values for CO (vol %), HC (ppm), and NO (ppm) emissions by vehicle MY are shown in Figures 6-8. Vehicles of MY 1980 and older were grouped together because very few measurements were available for older MY at each site. With respect to CO and HC emissions, the irregular values in vehicles from MY 1980 to 1990 with no control systems for emissions or A/F ratio are apparent. CO and HC emissions values started to decrease from 1988 MY

vehicles, which could be related to the adoption of the first regulated emissions standards for new vehicles at that time. A more steady decrease is observed in 1992 MY models, the first with two-way catalytic converters, and later in 1993, the year in which three-way catalytic converters were introduced. On average, before 1990 MY (30% of the total fleet) emits 2.6 vol % of CO, 850 ppm of HC, and 1394 ppm of NO. Furthermore, when one examines exclusively the 1993-2000 MY vehicles, the fleet averages are 0.50 vol % of CO, 201 ppm of HC, and 627 ppm NO. The decline in average emissions is positively very important; however, it remains high when compared with those of the 1983-1997 MY fleet of Chicago, which averaged 0.45 vol % CO, 214 ppm HC, and 409 ppm NOx, recorded in 1997 (21). A steady decrease in NO emissions was also noted, starting with the 1993 MY vehicles up to the 2000 MY, but recent models had low mileage accumulated at the time of the RSD measurements, i.e., degradation of the emission control system, if any, was not evident yet. Lean conditions on carbureted vehicles are noted by increase of NO emissions up to year 1990. Moreover, in Table 4, the mean emissions VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

399

FIGURE 8. NO average emissions as a function of vehicle model year. Error bars represent (95% confidence interval for the average of 12 sites.

TABLE 4. Average Emissions by Type of Vehiclea privates no. of vehicles CO (vol %) HC (ppm propane) NO (ppm)

taxis

32 478

3 939

1.2 + 0.23 1.5 + 0.12 409 + 78 478 + 38

LDT 5 951

class 3 HDGT

Supporting Information Available

449

A list of all site locations. This material is available free of charge via the Internet at http://pubs.acs.org.

1.4 + 0.29 2.6 + 0.23 461+ 128 881 + 274

935 + 127 1450 + 261 904 + 118 965 + 158

a

Standard error computed from averages of 12 sites. LDT, lightduty transport. HDGT, heavy-duty gasoline transport.

TABLE 5. Comparison of Mean Emissions Recorded in Mexico and the United Statesa location

CO (vol %)

HC (ppm propane)

NO (ppm)

MAMC 1991b MAMC 1994c MAMC 2000 Monterrey, Mex. 1995c Chicago 2000d Los Angeles 2000e

4.3 2 1.3 1.7 0.26 0.5

2100 890 440 670 94 120

nd nd 914 1430 316 420

a

nd, not determined.

b

Ref 11. c Ref 12.

d

Ref 21. e Ref 15.

of the matched vehicles this time grouped by the type of activity: private cars, taxis, light-duty trucks (LDT), and heavy-duty gross trucks (HDGT) are displayed. Noteworthy, NO emissions in taxis are the highest among the vintage of vehicles, a matter of concern if one takes into account that according to the millions of kilometers traveled per year, they represent the 22% of the total activity in the MAMC (5). Finally, the average emissions in the MAMC for the last 10 yr, monitored with the aid of RSD, are presented in Table 5. Although many factors are particular for each location, for example, load on vehicle, instrument calibration atmospheric variables, and socioeconomic, one can compare the behaviors of means of separate measurements. If one compares the three sets of data, a steady decrease in CO and HC emission is observed, but MAMC emissions are between 2 and 4 times higher than the averages reported for U.S. cities ap400

9

proximately at the same period. Before 1990, Mexican standards for new vehicles were lenient enough to be met without the employment of any emission control technology. The standards for 1991-1992 MY represented a transition period in which unleaded gasoline was available for the first time, and many of the vehicles introduced at that time (48%) had oxidative catalysts instead of TWC. From 1993 MY, all brand new cars in Mexico were brought under Tier 0, prevailing in the United States since 1981, once unleaded gasoline averaging 490 ppm sulfur was mandated for the MAMC. Furthermore, Tier I standard is requested for 2001 new vehicles, and a durability certificate for the emission control system will be imposed to gasoline vehicle makers after 2002 MY, but the sulfur content in the MAMC gasoline (400-500 ppm) and much higher outside of the zone (800 ppm) remains an important issue to resolve. The most important factor affecting any regional or seasonal sulfur control program is the need for the sulfur effect to be temporary. Vehicles are operated during all seasons and often travel across state lines. Thus, under essentially any regional or seasonal program, many or all vehicles in ozone nonattainment and maintenance areas would be occasionally or regularly fueled with high sulfur gasoline. However, there is concern that once a vehicle (particularly a new, less polluting vehicle) has been temporarily exposed to high sulfur levels, the negative emission impact may not be reversed when subsequently operated on gasoline with lower sulfur levels. Adopting stricter regulations in the MAMC should not be delayed if improvement of air quality is pursued.

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 37, NO. 2, 2003

Literature Cited (1) Programa para Mejorar la Calidad del Aire en el Valle de Me´xico, 1995-2000. Departamento del Distrito Federal, Gobierno del Estado de Me´xico, Secretarı´a de Medio Ambiente Recursos Naturales y Pesca, Secretarı´a de Salud, 2000. (2) The California Air Resources Board. Electronic Page, www. arb.ca.gov/msei/doctable.htm. (3) Code of Federal Regulations, Title 40, Part 86, Office of the Federal Register; U.S. Government Printing Office: Washington, DC, 1993. (4) Programa para Mejorar la Calidad del Aire de la Zona Metropolitana del Valle de Me´xico 2002-2010; Metropolitan Environmental Commitee, 2002 (in Spanish). (5) Statistics from the Departamento del Distrito Federal, Me´xico. Electronic Page, www.ddf.gob.mx/sedeco/transporte/ vehic.htm. (6) Bishop, G. A.; Stedman, D. H. Acc. Chem. Res. 1996, 29, 489. (7) Popp, P. J.; Bishop, G. A.; Stedman, D. H. J. Air Waste Manage. Assoc. 1999, 49, 1463. (8) Baum, M. M.; Kiyomiya, E. S.; Kumar, S.; Lappas, M.; Kapinus, V. A.; Lord, H. C., III. Environ. Sci. Technol. 2001, 35, 3735. (9) Guenther, P. L.; Bishop, G. A.; Peterson, J. E.; Stedman, D. H. Sci. Total Environ. 1994, 146/147, 297. (10) Bradley, K. S.; Brooks, K. B.; Hubbard, L. K.; Popp, P. J.; Stedman, D. H. Environ. Sci. Technol. 2000, 34, 897. (11) Beaton, S. P.; Bishop, G. A.; Stedman, D. H. J. Air Waste Manage. Assoc. 1992, 42, 1424. (12) Bishop, G. A.; Stedman, D. H.; De la Garza, J.; Davalos, F. J. Environ. Sci. Technol. 1997, 31, 3505. (13) Bohren, L. Profile of Vehicles at the Border Crossings between Tijuana, Mexico and San Diego, California. Proceedings of the Air & Waste Management Assocication, 93rd Annual Conference & Exhibition, Salt Lake City, UT, June 18-22, 2000 (CD version). (14) Gunderson, J. Memorandum: Mexico Border Study Progress Report; Air Resources Board: El Monte, CA, July 1997. (15) Pokharel, S. S.; Bishop, G. A.; Stedman, D. H. On-Road Remote Sensing of Automobile Emissions in the Los Angeles Area. Year 2; Prepared for the Coordinating Research Council, Inc.: GA, February 2001; CRC Project E-23-4.

(16) McClintock, P. M. The Colorado Enhanced I/M Program 0.5% Sample Annual Report; Prepared for The Colorado Department of Public Health and Environment: Tucson, AZ, January 27, 1998. (17) Cernuschi, S.; Guigliano, M.; Cermin, A.; Giovannini, I. Sci. Total Environ. 1995, 169, 175-183. (18) Biggs, D. C.; Akcelik, R. ITE J. 1986, 56, 29. (19) Tong, H. Y.; Hung, W. T.; Cheung, C. S. Atmos. Environ. 1999, 33, 2323. (20) Jime´nez, J. L.; Koplow, M. D.; Nelson, D. D.; Zahniser, M. S.; Schmidt, S. E. J. Air Waste Manage. Assoc. 1999, 49, 463.

(21) Popp, P. J.; Bishop, G. A.; Stedman, D. H. On-Road Remote Sensing of Automobile Emissions in the Chicago Area: A Progress Report. Proceedings of the 9th CRC On-Road Emissions Workshop, San Diego, CA, April 19-21, 1999. (22) Sjo¨din, A. J. Air Waste Manage. Assoc. 1994, 44, 397.

Received for review June 10, 2002. Revised manuscript received October 28, 2002. Accepted November 5, 2002. ES0207807

VOL. 37, NO. 2, 2003 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

401