Critical evaluation of US on-road vehicle emission inventories

Abstract

US Environmental Protection Agency estimates of on-road vehicle emissions are compared with ambient measurements and a fuel-based emission inventory. Several significant weaknesses and strengths are identified. (1) The emission estimates have varied considerably over the past 15 years and are not clearly converging to progressively more accurate and certain results. (2) The most recent emissions estimate accurately captures the rapid decrease in carbon monoxide (CO) and volatile organic compounds (VOC) emissions, but overestimates the magnitude of CO emissions by about a factor of two. (3) The oxides of nitrogen (NO x ) emission estimates for the mid to late 1990s are reasonably accurate, but NO x emissions have increased through that decade rather than decreased as indicated in emission estimates. (4) The most recent emissions estimate more accurately apportions NO x emissions between diesel and gasoline fueled vehicles than did earlier reports. (5) The ratio of two specific VOC species that has been characterized by ambient measurements suggests that the inventory speciation of the VOCs is inaccurate by factors of 3–4. These tests lead to the derivation of “inferred emissions” for CO and NO x from the US on-road vehicle fleet that are consistent with all information used in these tests. Finally, it is shown that the international picture of US emissions has significant inaccuracies and inconsistencies.