Neighborhood scale air quality modeling in Corpus Christi using AERMOD and CALPUFF
MetadataShow full item record
Ambient monitoring and air quality modeling of air toxics concentrations at the neighborhood-scale level is a key element for human exposure and health risk assessments. Since 2005, The University of Texas at Austin (UT) has operated a dense ambient monitoring network that includes both hourly automated gas chromatographs as well as threshold triggered canister samples and meteorological data in the Corpus Christi area. Although Corpus Christi is in attainment with the National Ambient Air Quality Standards for both ozone and fine particulate matter, its significant petroleum refining complex has resulted in concerns about exposure to air toxics. The seven site network, incorporating both the industrial and residential areas in Corpus Christi, provided a unique opportunity to further the development and understanding of air quality modeling for toxic air pollutants at the neighborhood-scale level. Two air dispersion models, AERMOD and CALPUFF, were used to predict air concentrations of benzene for one of the UT operated monitoring sites (Oak Park monitoring site: C634) and the predictions were compared to the observed benzene concentration data at the Oak Park monitoring site to evaluate model performance. AERMOD and CALPUFF were also used to predict benzene concentrations in populated areas and at sensitive receptor locations such as schools and hospitals. Both AERMOD and CALPUFF were able to reproduce the early morning high benzene concentration and the northern wind effect except under strong NNE wind conditions, where the observed data indicated elevated high benzene concentration which AERMOD and CALPUFF failed to predict. These under-predictions could be due to the NNE strong wind condition at that time of these occurrences or could be attributed to different types of emissions other than the point sources emissions from the 2005 TCEQ Photochemical Modeling inventory, such as mobile sources or accidental emission events. These preliminary analyses could be expanded by modeling longer periods, by including other emission sources and by inter-comparisons with observed data from other CCNAT monitoring sites. In addition, fundamentally different modeling approaches (eulerian, rather than lagrangian) could be considered.