Comparing Air Dispersion Model Predictions with Measured Concentrations of VOCs in Urban Communities

Abstract

Air concentrations of nine volatile organic compounds were measured over 48-h periods at 23 locations in three communities in the Minneapolis-St. Paul metropolitan area. Concentrations at the same times and locations were modeled using a standard regulatory air dispersion model (ISCST3). The goal of the study was to evaluate model performance by comparing predictions with measurements using linear regression and estimates of bias. The modeling, done with mobile and area source emissions resolved to the census tract level and characterized as model area sources, represents an improvement over large-scale airtoxics modeling analyses done to date. Despite the resolved spatial scale, the model did not fully capture the spatial resolution in concentrations in an area with a sharp gradient in emissions. In a census tract with a major highway at one end of the tract (i.e., uneven distribution of emissions within the tract), model predictions atthe opposite end of the tract overestimated measured concentrations. This shortcoming was seen for pollutants emitted mainly by mobile sources (benzene, ethylbenzene, toluene, and xylenes). We suggest that major highways would be better characterized as line sources. The model also failed to fully capture the temporal variability in concentrations, which was expected since the emissions inventory comprised annual average values. Based on our evaluation metrics, model performance was best for pollutants emitted mainly from mobile sources and poorest for pollutants emitted mainlyfrom area sources. Important sources of error appeared to be the source characterization (especially location) and emissions quantification. We expect that enhancements in the emissions inventory would give the greatest improvement in results. As anticipated for a Gaussian plume model, performance was dramatically better when compared to measurements that were not matched in space or time. Despite the limitations of our analysis, we found thatthe regulatory air dispersion model was generally able to predict space and time matched 48-h average ambient concentrations of VOC species within a factor of 2 on average, results that meet regulatory model acceptance criteria.

Indoor/ambient residential air toxics results in rural western Montana

“…Median concentrations of ambient air toxics measured in this program were two to eight times lower than those concentrations measured in urban areas such as Minneapolis, St. Paul (Pratt et al 2004;Adgate et al 2004) and Baltimore (Payne-Sturges et al 2004), 10-20 times lower than those measured in other urban areas such as Chicago, St. Louis (Sweet and Vermette 1992), and New Jersey (Wallace 1987), and ∼20-50 times lower compared to ambient air toxics measured in Los Angeles (Wallace 1987). Ambient concentrations of the BTEX compounds (benzene, toluene, ethylbenzene, and xylenes) measured in this study were comparable or slightly more elevated when compared to median concentrations reported in a four Oklahoma city study (Phillips et al 2005), and in Greensboro, NC (Wallace 1987).…”

Application, Evaluation, and Process Analysis of the US EPA’s 2002 Multiple-Pollutant Air Quality Modeling Platform

“…In particular, CMAQ-MP performs poorly for those short-live and highly active HAPs (e.g., 1,3-butadiene and acrolein), further investigation of the reactions associated with those species is warranted. Finally, the errors from measurements such as sample handling, accuracy of analytical standards, and a lack of site density may also contribute to the model biases, but the impacts of these factors are believed to be smaller as compared with other reasons [55].…”

Evaluation of Dispersion Model Performance in Predicting So2 Concentrations From Petroleum Refinery Complex

“…The RMSE is a measure of the deviations from the 1:1 relationship and preserves the scale of the original measurements. It is derived from the mean square error which comprised of bias (the extent of over or under estimation) and variance (precision) [8]. The fractional bias is presented because it is the statistic recommended by U.S.EPA.…”