| Motor vehicles are a major source of the air pollutants that damage the environment and affect human health and welfare. Emission rates from motor vehicles are a function of, among others, vehicle fleet technology characteristics (vehicle age, vehicle type, power fuel, gross vehicle weight, etc.). The MOBILE6 Vehicle Emission Modeling Software (the latest version of a software tool mandated by the EPA for predicting average in-use fleet emissions factors for HC, CO, and NOx) allows the user to utilize default values (estimated from 1996 nationwide data) or to input local fleet distributions. The accuracy of this input is important for two reasons. First, the fleet is composed of a wide variety of vehicles with different performance conditions, engine types, age, emission control equipment, fuel, etc. This means that the composition of the fleet can result in very different outcomes with respect to emissions. Second, unlike stationary sources, mobile motor vehicles are difficult to pinpoint because they can freely move around a non-attainment area. The emissions-producing characteristics of a fleet at a location at any given time might be very different at that location one hour later because of different types of vehicles now using the road network, or the vehicle fleet might be very different in one part of a metropolitan area as compared to others. For these reasons, MOBILE default values might not be appropriate for some modeling efforts, especially at finer levels of analysis. Modelers can approach these problems by generating fleet characteristics values which is customary done from regional vehicle registration data. This research found that, the locally registered fleet might not always represent the on-road fleet. The on-road fleet mix might be rather a function of time of the day and the surrounding site's characteristics.; This thesis used the statistical analysis of onroad observations from forty-four locations in the Georgia to develop a procedure to predict the fleet that is expected to be operated in an area using variables such as surrounding land uses, road functional class, time of the day, and the socioeconomic characteristics of the area (head of the household median income and age). These variables were used to develop an efficient method to determine sub-fleet mix (and the emissions-related characteristics of these vehicles) as a function of three levels of spatial aggregation—local, subregional, and regional—of the vehicle registration data. Research results provide the ability to better estimate fleet characterization input data, enhancing local and regional transportation-related air quality modeling. |
