| Petroleum substances and petrochemicals are widely used and there is a need to ensure their safe use. Risk assessments rely on the development of toxicity data and flexible methods to define the thresholds of no concern. The challenge for petroleum substances is in their compositional complexity. The constituents within petroleum substances have widely varying physicochemical properties, such as solubility, or vapor pressure and toxic potency. Further, in the aqueous environment, exposures can include both dispersed particulate oil as well as dissolved constituents. Environmental exposures are often transient, vary in time and space, such as during an oil spill, run off events, or from permitted outfalls. These variables create challenges in performing risk assessments since it is often difficult to extrapolate between lab and field, or even between substance types. Therefore, a combination of modeling and analytical tools have been developed and validated in the present study to provide a method for handling the complexities in the composition and exposures. The PETROTOX model provides an implementation of the hydrocarbon block method, which divides the chemical space of a petroleum substance into discrete "blocks" with relatively narrow range of physicochemical properties. This provides a flexible method for reading across between substances that is reflective of the composition and physicochemical properties of the constituents in the substance. In addition, a complimentary passive sampling analytical technique was developed to measure bioavailable hydrocarbons. The present work improves the technical basis for application of these tools and provides a case study application to a dispersed oil toxicity test. The last element of this dissertation research focuses on the expression of toxicity over time (e.g., toxicokinetics) following exposures to petroleum substances. Altogether, the methods and tools used in the present work provide the theoretical, and analytical, basis for evaluating risks of petroleum substances across many likely environmental exposure scenarios. |
