Impact of underestimating the effects of cold temperature on motor vehicle start emissions of air toxics in the United States.

ABSTRACT

Analyses of U.S. Environmental Protection Agency (EPA) certification data, California Air Resources Board surveillance testing data, and EPA research testing data indicated that EPA's MOBILE6.2 emission factor model substantially underestimates emissions of gaseous air toxics occurring during vehicle starts at cold temperatures for light-duty vehicles and trucks meeting EPA Tier 1 and later standards. An unofficial version of the MOBILE6.2 model was created to account for these underestimates. When this unofficial version of the model was used to project emissions into the future, emissions increased by almost 100% by calendar year 2030, and estimated modeled ambient air toxics concentrations increased by 6-84%, depending on the pollutant. To address these elevated emissions, EPA recently finalized standards requiring reductions of emissions when engines start at cold temperatures.

INTRODUCTION

Air toxics, which are also known in the U.S. Clean Air Act as "hazardous air pollutants" (HAPs), are those pollutants known or expected to cause cancer or other serious health and environmental effects. For example, some of these pollutants are known to have negative effects on human respiratory, cardiovascular, neurological, immune, reproductive, or other organ systems, and they may also have developmental effects. They may pose particular hazards to more sensitive populations, such as children, the elderly, or people with pre-existing illnesses. Mobile source air toxics (MSATs) are emitted by highway vehicles, non-road engines (such as lawn and garden equipment, farming and construction equipment, aircraft, locomotives, and ships), and their fuels. Air toxics are also emitted by stationary sources such as power plants, factories, oil refineries, dry cleaners, gas stations, and small manufacturers. Some MSATs of particular concern include benzene, 1,3-butadiene, formaldehyde, acrolein, naphthalene, polycyclic organic matter, diesel particulate matter, and diesel exhaust organic gases. (1) Benzene and 1,3-butadiene are both known human carcinogens. (2,3)

Future year emission estimates for air toxics are developed and air quality modeling is conducted by federal and state government agencies to set regulatory priorities and inform the decision-making process. Recent projected emission inventories and air quality modeling assessments by the U.S. Environmental Protection Agency (EPA) (4-6) predict that, with present and planned controls, light-duty highway vehicle air toxic emissions will decrease by approximately 70% between 1999 and 2020, and begin to increase slightly thereafter. For 49 states (other than California), these reductions are largely attributable to replacement of older technology light-duty vehicles with advanced-technology vehicles meeting more stringent emission standards under EPA's Tier 1, National Low-Emission Vehicle (NLEV), and Tier 2 vehicle emission control programs. All of these programs set emission standards at an ambient temperature of 75 [degrees]F. In California, reductions occur as a result of California Air Resources Board (CARB) emission standards, which are set at 50 [degrees]F as well as 75 [degrees]F. Until very recently, emission standards at lower temperatures were in place only for Colorado. As a result of work discussed in this paper, EPA recently set emission standards for volatile organic compounds (VOCs) at 20 [degrees]F. These standards will begin to phase in beginning in calendar year 2010.

When EPA developed its MOBILE6.2 emissions model, it assumed emissions occurring during vehicle starts at cold temperatures for advanced-technology light-duty vehicles and trucks would increase at the same rate relative to start emissions at 75 [degrees]F as for older vehicles. However, vehicle test data have shown that hydrocarbon emissions from vehicle starts for these advanced-technology vehicles are higher at cold temperatures than predicted by MOBILE6.2. Vehicle test data have been analyzed to estimate the increases as a function of temperature; these analyses have been documented, and are further described below. (1,7) Because the elevated hydrocarbon emissions are occurring at colder temperatures, there is little impact on projected ambient levels of ozone; however, implications for projected concentrations of air toxic components of VOCs are substantial in future years because these vehicles comprise a majority of the fleet. Although it is expected that particulate matter (PM) from vehicle starts would be similarly impacted by cold temperatures, the impact has not yet been estimated.

The purpose of this paper is to show the impact of underestimating air toxic vehicle emissions while starting in cold temperatures on estimated modeled ambient air concentrations in future years. This paper also shows the impact of EPA's recently finalized standards requiring a reduction in cold-temperature hydrocarbon emissions. The impacts of these cold-start emissions for PM are being evaluated.

METHODS

Analyses of Vehicle Emissions Test Data

To determine the effects of cold-temperature start on hydrocarbon emissions from advanced-technology light-duty vehicles and trucks, EPA analyzed data from vehicles and trucks meeting Federal Tier 1, Federal NLEV, Federal Tier 2, and California Low-Emission Vehicle standards. (7) Data were obtained from the following sources:

* vehicle emission certification data submitted by vehicle manufacturers to EPA as part of requirements to comply with requirements for cold-temperature carbon monoxide (CO) standards;

* surveillance testing data from the CARB; and

* test data collected by EPA at Southwest Research Institute (SwRI). (8)

These data were used to adjust the temperature and engine start emission factors for hydrocarbons in MOBILE6.2. Although data for almost 2000 vehicles were available in the EPA certification database, and data for 98 vehicles were available from CARB surveillance data, the data from SwRI were obtained from only four vehicles. However, the SwRI data provided the sole basis for estimating impacts at 0 [degrees]F.

Currently, MOBILE6.2 relies on multiplicative adjustment factors applied to basic emission rates to account for effects of cold-temperature starts on hydrocarbon emissions. These factors are derived for each segment, b, of the Federal Test Procedure (FTP):

TCF(b) = EXP[TC(b)*(T-75)] (1)

where TCF(b) is the temperature correction factor for an individual segment of the FTP(b), TC(b) is a correction factor coefficient that varies by model year, and T is temperature. (9) The correction factor coefficients are based on test data from 1983-1990 model year vehicles, collected by EPA, Environment Canada, and other agencies, but applied to modern, advanced-technology vehicles and trucks as well.

We modified the model to replace these multiplicative factors in properly functioning Tier 1 and later vehicles and trucks with additive values that were applied to 75 [degrees]F start emission factors. (1,7) These factors varied with temperature and vehicle technology (i.e., Tier 1, NLEV, Tier 2, etc.). Additive values can more closely approximate the additional hydrocarbon emissions caused strictly by the start and warm-up of the engine and/or the exhaust after treatment at the different temperatures than multiplicative values can. These values were obtained from subtracting the FTP emissions at 75 [degrees]F from the FTP emissions at 0, 20 and 50 [degrees]F using the data described above. Table 1 provides the additive adjustments that were used. The relationship between hydrocarbon emissions at 75 [degrees]F and lower temperatures in malfunctioning or deteriorated vehicles is not clear. Emissions could go up proportionally to properly operating vehicles or could go up at a lower rate. Although MOBILE6.2 currently uses a multiplier to account for temperature effects, doing so in the case of Tier 2 high-emitting vehicles leads to extremely high and unrealistic emission rates. Therefore we used the MOBILE6.2 estimate of FTP emissions for model year 2005 high-emitting vehicles at 20 [degrees]F in calendar year 2005 to develop an additive factor for all Tier 2 high-emitting vehicles. Although PM emissions are presumed to also be elevated from these vehicles at cold temperatures, EPA has yet to develop a model that includes these effects.