U.S. onroad transportation CO2 emissions analysis comparing highly resolved CO2 emissions and a national average approach : mitigation options and uncertainty reductions

The transportation sector is the second largest CO₂ emitting economic sector in the United States, accounting for 32.3% of the total U.S. emissions in 2002. Within the transportation sector, the largest component (80%) is made up of onroad emissions. In order to accurately quantify future emissions and evaluate emissions regulation strategies, analysis must account for spatially-explicit fleet distribution, driving patterns, and mitigation strategies. Studies to date, however, have either focused on one of these three components, have been only completed at the national scale, or have not explicitly represented CO₂ emissions instead relying on the use of vehicle miles traveled (VMT) as an emissions proxy. We compare a high resolution onroad emissions data product (Vulcan) to a national averaging of the Vulcan result. This comparison is performed in four groupings: light duty (LD) and heavy duty (HD) vehicle classes, and rural and urban road classes. Two different bias metrics are studied: 1) the state-specific, group-specific bias and 2) the same bias when weighted by the state share of the national group-specific emissions. In the first metric, we find a spread of positive and negative biases for the LD and HD vehicle groupings and these biases are driven by states having a greater/lesser proportion of LD/HD vehicles within their total state fleet than found from a national average. The standard deviation of these biases is 2.01% and 0.75% for the LD and HD groupings, respectively. These biases correlate with the road type present in a state, so that biases found in the urban and LD groups are both positive or both negative, with a similar relationship found between biases of the rural and HD groups. Additionally, the road group bias is driven by the distribution of VMT on individual road classes within the road groupings. When normalized by national totals, the state-level group-specific biases reflect states with large amounts of onroad travel that deviate significantly from the national average. We calculate the state-specific uncertainty of the Vulcan onroad emissions as a fraction of the state total emissions for each of the three sources; VMT, Age Distribution, and Fuel Efficiency. Uncertainty is largest for LD vehicles and urban roads that display more irregular and fuel-consuming start-and-go driving patterns. Therefore, states with greater urbanization levels (eg. New Jersey) and a larger proportion of LD vehicles (eg. California) generally display the largest levels of combined uncertainty. The disparity between expected and real emissions reductions, were policy to neglect spatial differences, highlights the importance of emissions mitigation strategies that incorporate the unique characteristics of geography in order to achieve consistently effective mitigation. In order to have measurable impact, mitigation must also ensure that potential reductions exceed the uncertainty associated with quantifying emissions. Thus climate agreements that fully account for uncertainties in emission estimates as well as regional differences will be best suited to enact the most effective policy.