A comparative study of US EPA 1996 and 1999 emission inventories in the west Gulf of Mexico coast region, USA.

Emission inventory is one of the required inputs to air quality models. To assist in the urban and regional modeling efforts, United States Environmental Protection Agency (EPA) has compiled a National Emission Inventory (NEI) for criterion pollutants, and the precursors of ozone and particulate matter (PM). In December 2002, EPA released the 1999 NEI estimates (NEI99), which represent the most recent national emission data. However, the data sets are not in model-ready format for air quality simulations. This present work converts the NEI99 Final Version 2 data sets into Inventory Data Analyzer (IDA) format and processes the data using the Sparse Matrix Operator Kernel Emissions (SMOKE) modeling system to generate a gridded emission inventory in a domain covering the west Gulf Coast Region, USA. The spatial and diurnal emission characteristics of the gridded emission inventories are then assessed and compared with those of the National Emission Trend 1996 (NET96). The NEI99 database contains more complete emission records in both area and point sources. It is also found that NEI99 data exhibit greater emissions with respect to point and mobile sources but smaller emissions with respect to area sources when compared to the corresponding gridded NET96 data in the same study domain. The most distinct differences between the NEI99 and NET96 databases are CO emission of mobile sources, SO2 emissions of point sources, and VOC/PM/NH3/NOx emissions of area and non-road sources. The gridded NEI99 data show low VOC/NOx ratios (<2-5) in the urban areas of the study domain.

Sensitivity analysis of ground-level ozone concentration to emission changes in two urban regions of southeast Texas.

Air pollutant emission is one of the predominant factors affecting urban air quality such as ground-level ozone formation. This paper assesses the impact of changing emission inventory scenarios, based on combinations of point, mobile, area/non-road and biogenic sources, on the tropospheric ozone concentration in two southeast Texas urban areas, i.e. Houston-Galveston and Beaumont-Port Arthur, during the rapid ozone formation event (ROFE) on August 25, 2000. The EPA's Community Multiscale Air Quality (CMAQ) modeling system with 1999 national emission inventory (NEI99) estimates and updated SAPRC99 chemical mechanism are used in the sensitivity analysis for twelve different emission scenarios. Based on model results, it is found that the point source emission of NOx and VOC contributes the greatest ozone peak in the ROFE. Removing Texas point sources of VOC and NOx emission from the inventory results in a reduction in peak O3 concentration by 128 and 70 ppbv in Houston urban area, respectively. Similar but less drastic impact from point source is also observed in the Beaumont-Port Arthur area. The effect on peak ozone concentration due to mobile, area and non-road sources emissions are less significant compared to that of point source emission. Reducing VOC emission appears to be more effective than reducing NOx emission in lowering peak O3 concentration in the studied region. Although biogenic emission can contribute up to 37 ppbv of peak ozone level over a large area, the affected area is away from the urban region of concern, and should not be the main cause for O3 non-attainment in the two urban areas. Removing CO emission from mobile sources does not lead to significant reduction (< 1 ppbv) in ozone concentrations. The modeled data also show that the transport of O3 precursors from adjacent states can cause a significant ozone plume near Beaumont due to its proximity to the state border based on the conditions during the August 25, 2000 O3 episode.