Evaluation of BME Data Fusion Approaches to Estimate SBTEX and to Inform Epidemiological Analyses in the US Gulf States

The United States Gulf States are home to many industries emitting styrene, benzene, toluene, ethylbenzene, and xylenes (SBTEX). Presently, adverse health effects of ambient exposure to SBTEX in highly polluted regions, like the Gulf States, must be evaluated. Epidemiological work, however, is limited by poor estimates of ambient SBTEX concentrations. Using Bayesian Maximum Entropy, several SBTEX estimation methods of varying resource intensity were evaluated, including observational interpolation (least intensity), incorporation of observational and emissions data trends (moderate intensity), and data fusion of observed and Comprehensive Air Quality Model (CAMx) data (most intensity). Generally, as method resource intensity increased, so did SBTEX estimation performance. Data fusion of observed and CAMx data was recommended as the best estimation method for ambient SBTEX in the Gulf States. Future epidemiological work, however, should consider differences in estimation uncertainty and the possibility of exposure misclassification for participants living within versus outside the CAMx-modeled domain