Photochemical Modeling of the Impact of Fuels and Vehicles on Urban Ozone Using Auto/Oil Program Data ALAN M. DUNKER* Environm ental Research Departm ent, General Motors Research and Development Center, 30500 Mound Road, Warren, Michigan 48090-9055 RALPH E. MORRIS † AND ALISON K. POLLACK † System s Applications International, 101 Lucas Valley Road, San Rafael, California 94903 CHARLES H. SCHLEYER Research Department, Mobil Research and Developm ent Corporation, 600 Billingsport Road, Paulsboro, New Jersey 08066-0480 GREG YARWOOD † System s Applications International, 101 Lucas Valley Road, San Rafael, California 94903 An extensive set of emission tests has been conducted in the Auto/Oil Air Quality Improvement Research Program on different fuel/vehicle systems. These emis- sion tests have been used to model the impact of fuel/ vehicle changes on ozone formation in Los Angeles, Dallas-Fort Worth, and New York in 1995 and 2005/2010. Light-duty vehicles are estimated to contribute 28-37% of the peak ozone in 1980/1985, decreasing to 7-18% in 1995, and further decreas- ing to 5-9% in 2005/2010. Gasoline changes that show promise in reducing the contribution of light-duty vehicles to ozone formation are reductions in olefin content,90% distillation temperature,sulfur content, and vapor pressure. Results for a methanol/gasoline blend (M85) used in prototype flexible/variable fuel vehicles depend on the assumptions used to project future M85 emissions. A research test gasoline produced less ozone than the M85 cases in Los Angeles and New York and either more or less ozone than M85 in Dallas-Fort Worth, depending on the assump- tions. Sensitivity tests for Los Angeles addressed uncertainties in the overall magnitude of emissions from light-duty vehicles, in the biogenic inventory, and in the representation of the atmospheric chemistry. Introduction Vehicles and fuels constitute a system, and the emissions from this system are determined by both the vehicle technologyand the propertiesofthe fuels. The fuel/vehicle system is a source of hydrocarbon, nitrogen oxide (NO x ), and carbon monoxide (CO) emissions, and, through a complex set of chemical reactions, these emissions con- tribute to ozone formation. Though emissions from the fuel/vehicle system and other sources in the U.S.have been controlled and reduced over the past 20 years, ozone concentrationsin 43urban areasremain above the National Ambient Air Quality Standard, based on monitoring data through 1993 (1). Consequently, additional emission controls have been instituted in recent years and more are being considered. Some of these controls focus on optimizingvehicle technologyforreducingemissions,some focus on changing the composition of gasoline, and some focus on introducing alternative fuels, such as methanol. The Auto/Oil Air Quality Improvement Research Pro- gram (AQIRP) was begun in 1989 as a cooperative research program among the three domestic U.S. auto companies and fourteen petroleum companies. The objective of the AQIRP is to develop data on potential improvements in vehicle emissions and air qualitysprimarily urban ozone levelssfrom reformulated gasoline,variousalternative fuels, and developments in automotive technology (2). The AQIRP has conducted over 3000 emission tests in two separate phases using different sets offuels and light-duty vehicles (cars and light-duty trucks). These tests are being analyzed to determine the effects of fuel/vehicle changes on the mass and composition of emissions. Information on the AQIRP results from the Phase I tests is available in an extensive seriesofreportsand publicationsthat describe the scope and limitations of the data (3-5). Photochemical modeling studies have been conducted previously to estimate the impact of alternative fuels on urban ozone, using various emission measurements and assumptions (6-10). The AQIRP data base for alternative fuels,however,is much larger and has more detailon more categories of emissions than data bases from prior testing programs. In addition, there are extensive AQIRP tests for gasolines with varying properties used in different fleets of vehicles. Some ofthe AQIRPdata have been used in another studyin which single-daysimulationswere conducted using a single-cell trajectory model (11). The study employed input data corresponding to average urban conditions, examined the ozone impactsforone test matrixofgasolines and a methanol/gasoline blend, and investigated the sensitivity of the ozone impacts to changes in the input data and the chemical mechanism. This paper presents results of detailed air quality simulations using data from the AQIRP Phase I tests and some initial Phase II tests. Multiday simulations were conducted, using a three-dimensional grid model, for gasolinesin severalAQIRPtest matricesand for a methanol/ gasoline blend. Vehicle emission inventories specific to three cities were developed and used in the modelingalong with other input data representing historical ozone epi- sodes. To the extent allowed by the AQIRP data and other information,fuel/vehicle effectswere estimated separately for the different categories ofvehicle emissions. Estimates are presented for the contribution oflight-duty vehicles to emissions and urban ozone in recent and future years, the effects of fuel/vehicle changes in future years, and the *To whom correspondence should be addressed;e-mail address: adunker@cmsa.gmr.com. † Present address: ENVIRON International Corp., 101 Rowland Way, Novato, CA 94945-5010. Environ. Sci. Technol. 1996, 30, 787-801 0013-936X/96/0930-0787$12.00/0  1996 American Chemical Society VOL. 30, NO. 3, 1996 / ENVIRONMENTAL SCIENCE & TECHNOLOGY 9 787