Top-down Tests of the U.S. On-Road Emission Inventory

Our laboratory and other workers have developed several “top-down” tests of emission inventories. These are tests conducted outside the structure of the emission inventory, and are based upon ambient concentration measurements and other independent information. They do not explicitly consider the individual components (emission factors, activity factors, etc.) that go into the development of inventories from the “bottom-up”. For the recent NARSTO Emission Inventory Assessment, we have reviewed published results of top-down tests of U.S. on-road vehicle emissions, and have extended some of them. These tests indicate that on-road vehicle emission inventories have some serious shortcomings. Specifically, the 2004 Trends Tables based upon MOBILE6 model: • Accurately capture the temporal trend of CO emissions, but overestimate the magnitude of these emissions by about a factor of 2. • Accurately estimate NOx emissions for recent years, but indicate a decreasing temporal trend through the 1990's, while top down evaluations (and previous Trends Reports) indicate an increasing temporal trend over that period. • More accurately reproduces the apportionment of NOx emissions between gasoline and diesel powered vehicles than did previous Trends Reports. • Accurately capture magnitude and temporal trend of VOC emissions. However, application of the SPECIATE software to national inventories indicates that the speciation of the VOCs may be inaccurate by factors of 3 or more.