Auto Asthma Index
Auto Asthma Index values range from 1 to 10, with 10 being the most polluting. The cutpoints separating the different Index values are the same that EPA uses to rate vehicles in their Green Vehicle Guide. [EPA (1) 2006]
The Auto Asthma Index for a vehicle depends upon the region in which the vehicle is driven. Smog forms through complex chemical interactions, such that different regions respond differently to the two different smog chemical pollutants, volatile organic compounds (VOCs) and nitrogen oxides (NOx). Vehicles are evaluated based on VOC emissions alone in regions where VOCs play the dominant role in smog formation. In regions where NOx is most important for smog control, the Auto Asthma Index is based on NOx emissions alone. In areas where both VOCs and NOx contribute to smogginess, total VOC and NOx emissions are used to rate vehicles. As a result, the Auto Asthma Index for a particular model varies with location. (More information about regional variation in smog formation is provided below.)
The majority of EPA's vehicle certification laws and state Smog Check programs allow trucks, SUVs, vans, and minivans to pollute more than cars. Our analysis breaks out of this tradition, holding all passenger vehicles to the same standard. Auto Asthma Index ratings are based on the least polluting technologies that are commonly put in today's cars. Several more powerful techniques for combating pollution exist but are not in widespread use in today's vehicles. Vehicles with a score of 1 are employing cleaner technologies, and those with a score of 10 are the worst polluters, either because they were built at a time with lax emissions standards, are in poor condition, or are modern vehicles that are massive or poorly designed.
We provide Auto Asthma Index values for vehicles in two different age ranges. Newer vehicles include those manufactured between 2001 and 2006. Older vehicles include those manufactured between 1985 and 2000. Auto Asthma Index values for vehicles in the two different age ranges should not be compared because they are derived from very different datasets. (More information about the the differences between the two datasets is provided below and in the FAQ). However, our analysis indicates that newer cars and trucks are typically better than older versions, due to tighter emissions standards, improvements in technology, and reduced wear and tear.
Data for the Auto Asthma Index came from two sources. Calculations for model year 1985 to 2000 vehicles were derived from California's Smog Check program, which requires biennial inspection for all registered vehicles that are at least six years old and were manufactured after 1975.
Newer vehicles are not subject to Smog Check, so to score model year 2001 and newer automobiles we used EPA test data and the EPA-certified maximum allowable pollution estimates vehicles when they reach midlife.
Our final Auto Asthma Index dataset contains 2.5 million individual records representing 10,102 unique vehicles.
Smog Check data (1985 - 2000 vehicles):
In order to test the real world pollution emissions from vehicles on the road today, EWG purchased a dataset containing emissions measurements from millions of vehicles tested in California's Smog Check program. The Smog Check program is a vehicle emissions monitoring program that includes a biennial test in which most autos are hooked up to a treadmill type device. Pollutant emissions measured during the course of this test can be translated into emissions in grams per mile driven using a vehicle's estimated fuel economy and some conversion factors. Several classes of vehicles, including diesels, some hybrids, extremely large vehicles, and all-wheel drive vehicles, are not subject to the particular type of Smog Check test we analyzed.
We analyzed Smog Check records for those vehicles certified to achieve California emissions standards, which are cleaner than federal standards. Because these cleaner vehicles were also using relatively clean California gasoline, and operating in a state with a well-established vehicle inspection and maintenance program designed to catch cars and trucks emitting excessive pollution, they are likely to be much less polluting than vehicles driven in many other states.
Smog Check data details:
EWG purchased files containing the results of all vehicles that received a Smog Check test during January, April, July, and October of 2004 from California's Bureau of Automotive Repair. By examining records from throughout the year, we controlled for the effects of seasonal gas formulations on vehicle emissions levels. We examined the test known as ASM 2525. An ASM test involves an emissions test on a treadmill type device (called a dynamometer), and forces the engine to put out power as it does when driving in the real world. The vehicle is run at 25 miles per hour, and its engine puts out approximately 25 percent of the maximum power required during the test EPA uses to estimate the fuel economy and emissions of a vehicle.
The ASM 2525 test is required biennially for vehicles in high ozone regions of California that are at least six years old and were manufactured after 1975. All vehicles with diesel or natural gas engines, those weighing more than 8,500 pounds when fully loaded (the gross vehicle weight rating), and those with non-disengageable traction control or all-wheel drive (full-time four-wheel drive), are exempt from this type of test, and are not included in our analysis.
We made every effort to eliminate faulty emissions measurements, created for example by broken testing equipment, from the Smog Check data. To this end, we purged vehicle records with negative emissions measurements, and with carbon dioxide measurements above 18 percent, combined carbon dioxide and carbon monoxide measurements exceeding 20 percent, oxygen levels over 20 percent, and NOx measurements greater than 20,000 parts per million, as recommended by researchers at the Lawrence Berkeley National Laboratory [Wenzel 2000]. We also removed records that listed a dilution correction factor of less than 0.9 or greater than 3.0, as this is an indication of problems such as leakage within the emissions testing equipment [Wenzel 2000]. Finally, we eliminated records for vehicles that were pre-tested, or repaired before their test (because they failed an earlier Smog Check test), and for vehicles whose tests were aborted.
The measurements of smog chemicals provided in the Smog Check records have been adjusted for relative humidity, as well as for the small amount of dilution of emissions that occurs during use of functional Smog Check testing equipment. However, because we reassigned measurements of zero parts per million of VOCs or NOx to 0.5 parts per million, a value representing half the detection limit for these pollutants, we had to adjust these new values for the relative humidity and dilution measured during the tests.
Pollution emissions from a vehicle typically decline over the 75 second duration of the Smog Check ASM 2525 test [Wenzel 2000]. California allows vehicles to "fast pass," or end the emissions test early, if a vehicle emits pollutants at concentrations below specific cutpoints during the test. This is because Smog Check tests are designed to flag only the worst polluters among vehicles of a particular age and weight. As a result, some passing cars and trucks might have artificially elevated emissions measurements, because their emissions were low enough to allow the test to end early.
Finally, VOC concentrations from Smog Check tests, measured as hydrocarbons, were adjusted to omit the non-reactive methane fraction of the organic gases that does not lead to smog formation. This simply requires multiplying hydrocarbon values by 0.984. [EPA 2004]
After cleaning and adjusting the emissions data, Smog Check records were grouped based on vehicle make, model, year, size class, transmission, engine size, and number of cylinders. Grouping of records was hindered by lack of standardization in the model name provided for each record by Smog Check technicians. Automated and manual matching techniques were used to group unconventionally-named vehicle records appropriately.
Some vehicles can be purchased with a variety of different engine and equipment options that could affect smog chemical emissions. However, since information about these options is not included in Smog Check records, we could not examine the effects that this equipment might have on vehicle emissions. If fewer than 20 records were available to describe a particular vehicle, we eliminated it from our Auto Asthma Index database. We also removed vehicles manufactured before 1985, as the Index values for these vehicles were uniformly abysmal.
We translated vehicle emissions measured using the Smog Check treadmill test to reflect grams of pollution emitted per gallon of gas burned using a formula derived by scientists at Lawrence Berkeley National Laboratories [Wenzel 2000]:
D1 = molar concentration of fuel carbon in exhaust
D1 = carbon monoxide + carbon dioxide + (6 x VOCs/10,000)
D2 = moles of carbon per gallon of California reformulated gasoline
D2 = 0.85 (g C / g gasoline) x 743 (g gasoline / L) x 3.8 (L / gal) / 12 (g C / mole)
VOCs (grams per gallon) = 2 x [(VOCs/10,000) / D1] x D2 x 86
NOx (grams per gallon) = [(NOx/10,000) / D1] x D2 x 46
where concentrations of VOCs and NOx are measured in parts per million, and concentrations of carbon monoxide and carbon dioxide are measured in percent. All of these measurements are provided by the Smog Check test.
We then used EPA's fuel economy estimates to express emissions as grams of pollution per mile traveled [EPA (2) 2006]. EPA calculates an overall fuel economy for vehicles using a harmonic average of the fuel economies measured based on tests designed to simulate city and highway driving [EPA 2005]. EPA assumes a typical car drives 55 percent of its miles in the city, and 45 percent on highways. Before calculating the harmonic average, EPA now adjusts the fuel economy of the tests to compensate for the well-known bias towards unrealistically high values, using just 90 percent of the measured city fuel economy, and 78 percent of the measured highway fuel economy. We used the same adjustment and averaging method to calculate overall fuel economies for all vehicles in our Smog Check dataset.
When multiple fuel economy estimates were available for a particular vehicle based on distinctions not recorded in the Smog Check data, we picked the highest one, in order to use the most conservative estimate of vehicle emissions. For example, the Smog Check records often do not specify whether vehicles are two- or four-wheel drive. We made the more conservative assumption that these vehicles were the more fuel efficient two-wheel drive versions. Using our conversion method, a vehicle with higher fuel economy produces fewer emissions than the same vehicle with lower fuel economy.
We were unable to match a small number of Smog Check records with an EPA fuel economy estimate. These records could not be used to calculate smog chemical emissions in grams per mile, and were eliminated from the dataset.
Our Smog Check dataset provided 2.5 million individual records representing 6,365 vehicles.
Smog Check data analysis:
The Auto Asthma Index value for each vehicle model was determined using the median emissions level for this model. As described above, for regions where smog formation is primarily controlled by VOC emissions, the Auto Asthma Index value was calculated based on median VOC emissions, while for regions where smog formation is primarily controlled by NOx emissions, the Index value was calculated based on median NOx emissions. In regions sensitive to both groups of smog chemicals, the Auto Asthma Index value was determined by adding the emissions of both smog chemical groups together for each individual vehicle, obtaining the median of this additive value for each vehicle model, and then using scoring cutpoints for total smog chemical emissions obtained by adding together the EPA values for the two pollutants individually. [EPA (1) 2005] In addition, we compared the median emissions values to the minimum emissions for the dirtiest five percent of the individual vehicles representing each model.
To determine whether a particular model was better or worse than similar, typical models, the model's median emissions level was compared to the median of median emissions levels for all models of the same age and vehicle type. Median emissions levels were also used to rank vehicles as one of the best or worst vehicles for a particular region. Again, median VOC emissions were analyzed for vehicles in VOC sensitive regions, while median NOx emissions were analyzed in NOx sensitive emissions. In regions sensitive to both chemical groups, the VOC and NOx emissions measured in each individual Smog Check record were added together, and this additive value was used to calculate the median total smog chemical emissions of each vehicle model, as well as the median of medians for all models of the same age and vehicle type. These additive smog chemical emissions values were used to judge vehicles as better or worse than similar models, as well as for ranking vehicles as among the best or worst for regions sensitive to both VOCs and NOx.
Smog Check Gross Polluter details:
"Gross Polluters" are vehicles that emit double or more the allowable pollution according to the California Smog Check program. [Shafizadeh 2004] The California Smog Check program applies emissions cutpoints to vehicles within specific age and size classes, so they do not excessively penalize older cars and trucks built with outdated emissions control technology and required to meet less stringent emissions standards. These cutpoints are much looser than the EPA certification standards for all vehicles, allowing cars and trucks to emit far more pollution than the amount certified by EPA as the maximum allowable level of pollution within the first 100,000 to 150,000 miles of use.
We calculated the percentage of Gross Polluters for each model in our Smog Check database. Older vehicles are more likely to be Gross Polluters because they have experienced more wear and tear, and may not have been maintained well. However, many vehicles are Gross Polluters due to poor construction on the assembly line, or use of substandard engines and emissions control equipment.
Auto Asthma Index for vehicles manufactured after 2000:
Newer vehicles are only required to have a Smog Check after a change of ownership. Therefore, we had very few Smog Check records for vehicles manufactured after 2000. We used the EPA certification values as the best estimate of pollution emissions. These values are the maximum allowable emissions from a vehicle once it has been driven 100,000 miles (for cars), 120,000 miles (for trucks) or 150,000 for the newest, cleanest vehicles (Super Ultra Low Emissions Vehicles or Partial Zero Emissions Vehicles).
EPA and automakers test vehicles using a standard series of tests and test fuels, as required by EPA's the Federal Test Procedure (FTP). The FTP test was designed to mirror real world driving conditions, but has been roundly criticized for being out of date with today's driving conditions. In a FTP test, a vehicle is driven an average speed of about 21 miles per hour, with a brief maximum of 57 miles per hour. The test is about 40 minutes. Emissions produced when the vehicle starts up, accelerates, and idles are all included. Emissions are expressed in terms of pollution per mile traveled. Automakers submit emissions test information to EPA, along with modeling assumptions to estimate vehicle emissions at mid-life. EPA performs its own tests on a fraction of vehicles.
EPA data details:
We obtained all FTP data for model year 2000 to 2006 vehicles tested by EPA from EPA's Certification Test Results website. [EPA (3) 2006] Data included vehicle manufacturer, division, model year, model name, engine characteristics (engine size, transmission), and the area of the country where the vehicle would be sold. The number of engine cylinders was not available in FTP datasets.
Only emissions test data for common fuels was included in our analysis. These include unleaded (EPA code 06), CARB Phase II Gasoline (23), and Tier 2 Unleaded (61). Diesel and alternative fuels were omitted. Manufacturers of prototype vehicles were also omitted. These include Baytech, GFI, IMPCO, and Quantum Technologies. Vehicles that were clearly designed for commercial use were also excluded.
Vehicles sold in California, Massachusetts, New York, Vermont, and Maine are required to meet more stringent emissions standards than vehicles sold in other states. However in all states, seemingly identical automobiles may be certified to emit varying amounts of pollution, and most drivers are unaware of the exact pollution certification of the vehicle they own. Automakers keep their overall fleet averages of emissions below specific regulatory thresholds by offering several versions of a particular model, balancing sales of more polluting versions with sales of less polluting ones. We gave a single Auto Asthma score for each vehicle by displaying the cleanest version of a vehicle sold in a particular state.
EPA data analysis:
Using methods designed by EPA's Green Vehicle Guide we calculated an Index for each model in a particular region of the U.S. [EPA (1) 2005] We grouped test records by manufacturer, carline name, engine size, transmission type and year, and calculated the minimum and maximum values of smog chemical emissions when multiple pollution levels existed. When state-specific test data are available, users are offered the Auto Asthma Index for typical cars sold in their state. However, when only California- or federally-certified vehicles were tested, that score is presented. When multiple versions of a federal or California vehicle were tested, the minimum pollution measurement was used to determine the Auto Asthma Index score. This gives the most conservative estimate of pollution emissions among a range of automobiles.
One exception to this scoring method was our treatment of the 160 vehicles tested for which some models are certified as Partial Zero Emissions Vehicles (PZEV) or Super Ultra Low Emissions Vehicles (SULEVs). PZEV and SULEV models exist for many common small vehicles (Honda Accord, Toyota Camry, Subaru Outback, and Ford Focus are examples). However, these vehicles make up a small proportion of the overall sales of these popular models. We flagged models for which a PZEV or SULEV version is available, but did not assign these models a score of 1, since a typical driver most likely does not own one of these rare, clean versions. For 16 models, 100 percent of the vehicles sold meet the PZEV standard. These few were assigned a score of 1 in the Index.
To calculate Auto Asthma Index values, we examined FTP tests for volatile organic compounds (VOCs) and nitrogen oxides (NOx), measured in units of grams per mile driven. VOCs were measured as non-methane organic gases (NMOG) only, eliminating the non-reactive methane gas that is not involved in smog formation. Where test data for both pollutants was available we added VOC and NOx values for an estimate of total smog precursor chemicals produced by each vehicle. Auto Asthma Index scores were based on VOC emissions only for VOC sensitive regions of the US, and NOx emissions only for NOx sensitive regions. In regions sensitive to both smog chemical groups, Auto Asthma Index values were calculated by summing expected emissions of the two smog chemical groups, and using scoring cutpoints for total smog chemical emissions obtained by adding together the EPA Green Vehicle Guide values for the two pollutants individually. [EPA (1) 2005]
EPA does not test each vehicle annually. Vehicles are more likely to be retested when there is a modification that will affect emissions or fuel economy. When test information was not available for a particular model, we extended EPA emissions estimates from the previous model year. The model year 2000 data were included in the database solely to provide information about 2001 model year vehicles, and were not used for subsequent analysis.
To determine whether or not a vehicle was actually produced and sold during the year following an emissions test, we downloaded EPA's Fuel Economy Guide [EPA (2) 2005] and matched the two files based on manufacturer, carline name, model year, engine size, transmission and state/federal certification. Vehicle names were often inconsistent between our two files. We manually scanned and matched some vehicle names, but many other models were unavoidably dropped from our analysis due to inconsistent naming between the two EPA data files.
Our analysis criteria limited FTP records to 2,997 unique automobiles with test data for at least one pollutant. We were able to extend FTP test records to the subsequent year for just 587 models. We provide Auto Asthma Index scores and comparisons for 3,737 vehicles.
EPA and automakers report both the maximum and likely emissions estimates for vehicles once they have reached mid-life. Likely estimates are the results from the emissions test. They are typically about half of the maximum allowable emissions. The difference between likely and maximum emissions gives automakers a margin of safety for individual vehicles that hold up poorly over time.
We analyzed a vehicle's likely and maximum estimates for different purposes. The Auto Asthma Index uses maximum allowable emissions and EPA's Green Vehicle Guide cutpoints [EPA (1) 2006] to score vehicles. For ranking and analysis of the best and worst vehicle models, we first sorted vehicles by their Auto Asthma Index scores (which are based on maximum estimates) and then ranked vehicles with the same scores by their likely emissions. The same ranking method was used to determine whether individual vehicles were above or below the median emissions value for all models of the same year and vehicle type.
Classification by EPA vehicle type:
EPA vehicle types were derived from EPA's Fuel Economy Guide for all vehicles. [EPA (2) 2006] Both Smog Check and EPA data were linked to the Fuel Economy Guide files downloaded from EPA, matching automaker, model name, year and vehicle characteristics. Vehicle names were not standardized and both automated and manual matching were used to link the files.
We considered the following types of vehicles: small, medium, and large cars, station wagons, trucks, SUVs, vans (both passenger and commercial), minivans, and oversized vehicles weighing more than 8,500 pounds fully loaded (gross vehicle weight rating). While EPA provides classifications of small, medium, and large station wagons, there were very few station wagons relative to passenger cars, so we did not maintain this distinction. We did not include cab chassis or other solely commercial automobiles. For vehicles manufactured before 1998, minivans designations were added in by hand, as the original records did not distinguish between vans and minivans.
High Miles Vehicle details:
Typically, the more miles a vehicle is driven, the more smog chemical pollution it produces, as the engine and emissions control equipment deteriorate with use. [Wenzel 1997] Air quality engineers in California have estimated the average number of miles a vehicle has been driven by model year, taking into account the fact that older vehicles are driven less than newer ones. [Kear 2003, citing EMFAC 2001] We identified high miles vehicles as those driven at least 10 percent more than this average number of miles for vehicles of the same model year.
Lax Emissions Requirements details:
We identified vehicles allowed to meet lax emissions requirements based on two criteria. 'Truck-type' vehicles, including trucks, SUVs, vans, and minivans, as well as oversized vehicles, those weighing more than 8,500 pounds fully loaded, are both allowed to pollute more according to current federal and California regulations. [NRC 2006] Vehicles in these categories were flagged as meeting lax emissions requirements.
In addition, less protective emissions standards in place before 1996 allowed vehicles produced before this time to emit more smog chemical pollution. [NRC 2006] Any vehicle manufactured before 1996 was flagged as meeting lax emissions requirements.
Best and Worst Automakers details:
The "Top Ten" automaker analysis presents the average Auto Asthma Index value for all vehicles manufactured by the ten largest parent companies in the American auto market between 2001 and 2006 (EPA data only). The brands consolidated under each of the parent companies are as follows: BMW: BMW, Hireus, Mini, Rolls-Royce; Daimler-Chrysler: Chrysler, Dodge, Europa, Jeep, Mercedes-Benz, Plymouth; Ford: Aston Martin, Ford, Jaguar, Land Rover, Lincoln, Lincoln-Mercury, Mercury, Roush Performance Products, Volvo; GM: Buick, Cadillac, Chevrolet, Daewoo, GMC, Hummer, Oldsmobile, Pontiac, Saab, Saturn; Honda: Acura, Honda; Hyundai: Hyundai, Kia; Mazda; Nissan-Renault: Infiniti, Nissan; Toyota: Lexus, Toyota; Volkswagen: Audi, Bentley, Lamborghini, Volkswagen.
We also calculated the average Auto Asthma Index value for individual divisions based on models manufactured during the same period, as well as for models manufactured between 1996 and 2000 (Smog Check data). (Read our automaker analysis.)
Regional variation in the Auto Asthma Index
This report focuses on some of the smoggiest regions of the U.S., according to data from ozone monitors analyzed by the American Lung Association [ALA 2004, 2005, 2006]. Examination of planning documents provided by local air quality agencies, as well as the scientific literature on regional atmospheric chemistry, reveals differences in the influence of the two smog chemical groups, volatile organic compounds (VOCs) and nitrogen oxides (NOx), over the formation of ozone in different areas. [Hidy 2000, Solomon 2000] Some regions are most sensitive to emissions of VOCs. This is typical of the heavily trafficked, dense urban cores of big cities. It is also often true of the dry Western U.S. Other regions are most sensitive to NOx emissions. This is typical of rural areas, and areas surrounded by forests or lots of vegetation. It is also typical of the Southeastern U.S. In other regions, neither chemical has a clearly dominant influence over the formation of ozone. These regions might be sensitive to NOx one day, and sensitive to VOCs the next. Or they might contain an urban core that is sensitive to VOCs, and a few surrounding suburbs that are sensitive to NOx. (See a list of the smoggiest regions of the U.S., and the pollutant primarily responsible for smog formation in each of these regions.)
Because different regions of the U.S. respond differently to emissions of the two smog chemical groups, we have provided Auto Asthma Index ratings for vehicles based on the most important smog chemicals for each region. In areas sensitive to VOCs, we have rated vehicles based on their emissions of VOCs; in areas sensitive to NOx, we have rated vehicles based on their emissions of NOx; and in areas where neither precursor chemical is dominant, we have rated vehicles based on their overall emissions of smog chemical pollutants. When ranking vehicles as best or worst for a particular region, VOC emissions only were used for VOC sensitive regions, NOx emissions only were used for NOx sensitive regions, and total smog chemical emissions were used for regions sensitive to both groups of chemicals.
When comparing individual vehicles driven within regions sensitive to either VOCs or NOx, we have assumed that the minute, incremental changes to emissions of the dominant smog forming chemical group are both positively and linearly related to similarly minute, incremental changes in smog. As a result, we can make a quantitative comparison of the effect of individual vehicles on air quality and public health. We cannot scale this assumption up to larger groups of vehicles, however, because larger changes to smog chemical emissions produce decidedly nonlinear changes to smog levels. [Solomon 2000]
We also cannot make the same assumption in regions where neither chemical group dominates the smog formation process. This is because a portion of NOx emissions can actually scour ozone, the primary component of smog, from the atmosphere locally. [Heuss 2003] Therefore, in regions where neither precursor is dominant, the NOx portion of the combined vehicle emissions could have a positive or negative effect on regional ozone, depending on a large number of factors that includes weather, time of day, and the specific location of the vehicle within the region. In this case, we can compare vehicles quantitatively concerning the amount of smog chemicals they emit, and qualitatively concerning their effect on air quality and public health.
Because we rely solely on VOC emissions to compare vehicles in VOC sensitive regions, we do not account for the possibility that a small portion of NOx emissions can reduce smog locally. NOx emissions themselves are associated with a number of air quality and public health issues, primarily as contributors to particulate matter pollution, offsetting any benefit derived from their capacity to scour ozone from the atmosphere. [WHO 2003]
American Lung Association. 2004. State of the Air: 2004. http://lungaction.org/reports/stateoftheair2004.html
American Lung Association. 2005. State of the Air: 2005. http://lungaction.org/reports/stateoftheair2005.html
American Lung Association. 2006. State of the Air: 2006. http://lungaction.org/reports/stateoftheair2006.html
California Air Resources Board. EMFAC 2001 V2.08. As cited in: Kear T, Niemeier D. 2003. Composite exhaust emissions rates: Sensitivity to vehicle population and mileage accrual assumptions. Transportation Research Record No. 1842, Energy, Air Quality, and Fuels 2003. Transportation Research Board of the National Academy of Sciences. Washington, DC. Available at: aqp.engr.ucdavis.edu/Documents/sensitivity.pdf
Environmental Protection Agency. 2004. Conversion Factors for Hydrocarbon Emission Components, NR-002b. Assessment and Standards Division, Office of Transportation and Air Quality, U.S. Environmental Protection Agency, EPA420-P-04-001. http://www.epa.gov/otaq/models/nonrdmdl/
Environmental Protection Agency. 2005. Light-duty Automotive Technology and Fuel Economy Trends: 1975 Through 2005. Office of Transportation and Air Quality, U.S. Environmental Protection Agency, EPA420-R-05-001. http://www.epa.gov/otaq/cert/mpg/fetrends/
Environmental Protection Agency. 2006. Green Vehicle Guide. http://www.epa.gov/greenvehicles/ (Cutpoints are available at: http://www.epa.gov/autoemissions/rating.htm, accessed 5/15/06)
Environmental Protection Agency. 2006. www.fueleconomy.gov. Data available for download at http://www.fueleconomy.gov/feg/download.shtml
Heuss JM, Kahlbaum DF, Wolff GT. 2003. Weekday/Weekend ozone differences: What can we learn from them? Journal of the Air and Waste Management Association 53:772-788.
Hidy GM. 2000. Ozone process insights from field experiments Part I: Overview. Atmospheric Environment 34:2001-2022.
Kear T, Niemeier D. 2003. Composite exhaust emissions rates: Sensitivity to vehicle population and mileage accrual assumptions. Transportation Research Record No. 1842, Energy, Air Quality, and Fuels 2003. Transportation Research Board of the National Academy of Sciences. Washington, DC.
National Research Council. 2006. State and Federal Standards for Mobile-Source Emissions. Committee on State Practices in Setting Mobile Source Emissions Standards, Board on Environmental Studies and Toxicology, Division on Earth and Life Studies, National Research Council of the National Academy of Sciences. National Academies Press. Washington, DC. Available at: http://darwin.nap.edu/books/0309101514/html (Accessed May, 2006)
Shafizadeh K, Niemeier D, Eisinger DS. 2004. Gross Emitting Vehicles: A Review of the Literature. Prepared for the California Department of Transportation, Task Order No. 27. Available at: http://aqp.engr.ucdavis.edu/Documents/Gross%20
Emitter%20Lit%20Review%20v11%5B1%5D.doc (Accessed May, 2006)
Solomon P, Cowling E, Hidy G, Furiness C. 2000. Comparison of scientific findings from major ozone field studies in North America and Europe. Atmospheric Environment 34:1885-1920.
Wenzel T, Ross M, Sawyer R. 1997. Analysis of emissions deterioration of in-use vehicles, using Arizona IM240 data. Presented at the Society of Automotive Engineering Government/Industry Meeting, May 5-7, 1997, Washington, DC. Available at: http://enduse.lbl.gov/projects/vehicles/appendix_N.pdf (Accessed May, 2006)
Wenzel T, Singer B, Gumerman E, Sawyer R, Slott R. 2000. Evaluation of the Enhanced Smog Check Program: A Report to the California Inspection and Maintenance Review Committee. Lawrence Berkeley National Laboratory, LBNL-46117. http://www.imreview.ca.gov/styles/smogcheck/
World Health Organization. 2003. Health Aspects of Air Pollution with Particulate Matter, Ozone and Nitrogen Dioxide. Report on a WHO Working Group. Bonn, Germany. http://www.euro.who.int/air/activities/20030612_1