Revision of Light-Duty Vehicle Emission Inventories Using Real-World Measurements-Auto/Oil Program, Phase II

Abstract

ABSTRACT Predictions of the models MOBILE5a and EMFAC7F were revised to more accurately represent emissions from in-use vehicles under real-world conditions. First, exhaust enrichment emissions were added to account for high accelerations that occur in about 2% of in-use driving but are not contained in the Federal Test Procedure. The added emissions were estimated from tests on an alternative driving cycle and a survey of driving patterns. Second, model predictions of hot soak emissions were adjusted using new data from a study of in-use vehicles. Lastly, model predictions of diurnal emissions and resting losses were adjusted using results from new three-day tests in which the temperature is varied in real time. The adjusted model predictions were used to develop emission inventories for Los Angeles, New York, and the Chicago-Milwaukee region. The net adjustments from all of the changes to hydrocarbon, CO, and NOx emissions in years 1987-1991 were less than 2%, 9%, and 1%, respectively, of the light-duty vehicle emissions. These net adjustments were relatively small because individual adjustments tended to cancel; because MOBILE5a and EMFAC7F predictions were combined with new real-world data when making the adjustments; and because adjustments were developed for only those categories of emissions for which new data were available. However, percentage adjustments for future years and for some subcategories of evaporative emissions were larger. Predictions of the two models were also compared before any adjustments. Unadjusted MOBILE5a predictions of hot soak emissions for a 1988 California fleet are 145% higher than unadjusted EMFAC7F predictions, and MOBILE5a predictions of hot stabilized exhaust emissions are 18-122% higher for a 1987 California fleet. Unadjusted MOBILE5a predictions of first-day diurnal plus resting-loss emissions are 26% lower than EMFAC7F predictions for carbureted vehicles with functioning emission control systems and 41% higher for fuel-injected vehicles in a 1988 fleet.