Description:

Bayesian Models for Analysis of Airborne Chemical Exposures During the Deepwater Horizon Oil Spill Response and Clean-up Efforts

My research focuses on developing statistical methods to quantify airborne chemical exposure in response to the Deepwater Horizon oil rig fire and other settings in environmental health. Factors complicating the exposure estimation include analytical method and data collection limitations. All analytical methods used to measure chemical concentrations have a limit of detection (LOD), or a threshold below which exposure cannot be detected with the analytical method (measurements below LOD are called censored measurements). However, even these small exposures must be assessed to provide the most accurate estimates of exposure. Similarly, due to the scope of this event, it was not possible to take measurements in all scenarios where workers were involved in the response.

In this talk, I describe a bivariate left-censored Bayesian model used to quantify exposures under possible LOD censoring in both the response and predictor. I also describe how this method can be expanded to a multivariate framework with multiple chemical predictors with possible censored observations. Then, I briefly describe how we used a database of over 26 million direct-reading VOC area measurements to supplement our exposure information for THC. Finally, I conclude with possible avenues for future research in environmental health and exposure assessment.