Date of Award

6-1-1999

Embargo Period

1-1-2025

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

Department

Biometry and Epidemiology

College

College of Graduate Studies

First Advisor

Eberhard O. Voit

Second Advisor

Karl J. Karnaky, Jr.

Third Advisor

Philip F. Rust

Fourth Advisor

Geoffrey I. Scott

Fifth Advisor

Zhen Zhang

Abstract

The first part of this dissertation presents an analysis of polycyclic aromatic hydrocarbon (PAH) sediment and oyster contamination data collected at Murrells Inlet, South Carolina. Chapters 2 and 3 present distribution analyses of the sediment and oyster contamination data using three representative PAH analytes: phenanthrene (PHE), pyrene (PYR), and chrysene (CRY). The results indicate that the Weibull gives an adequate fit for almost all the PAH analytes considered. In fact, the Weibull almost always provides a better fit to the data than the lognormal distribution. Chapter 3 also addresses issues associated with non-detection points, as they are regularly encountered in environmental analyses. Several statistical methods for estimating the Weibull parameters from such left-censored data are explored. The overall result is in agreement with recent findings reported by other investigators; methods based on the underlying distribution of the data give more consistent results than those obtained by commonly used substitution methods. In the second part of the dissertation, two competing models are presented that predict PAH bioavailability: the Equilibrium Partitioning (EqP) model and the Independence Uptake (IU) model. In Chapter 4, statistical tests and scatter plots of lipid and organic carbon-normalized data for PHE, PYR, and CRY are presented that clearly indicate the inappropriateness of EqP model point-estimated biota-to-sediment ratios (BSRs) for this system. As an alternative, the IU probability model is developed that predicts PAH uptake in situations where the EqP model is not appropriate. The IU model assumes that the sediment and oyster PAH concentrations are independent at a given sampling site, and treats the ratio of contaminant concentrations between these two phases as a random variable with a corresponding probability distribution. These probability models give 'weight' to the BSR data whereby some values are more likely to occur than others. Contrary to the predictions of the EqP model, BSR values under the IU model appear to decrease with an increase in the molecular weight of the analyte. This suggests that heavier PAHs are 'falling out' of the water column into the sediment and become less available for uptake by oysters. In Chapter 5, the methods developed in Chapters 2, 3, and 4 are applied to all the PAH analytes collected at Murrells Inlet. Finally, the IU model developed for the Murrells Inlet data is applied to sediment data collected at Shem Creek, Charleston Harbor, in order to estimate possible PAH concentrations distributions in oysters for this estuary. Using Monte Carlo techniques, BSR values were randomly chosen based on the analyte-specific probability distributions of the BSR data. These simulations were repeated numerous times in order to create an ensemble of possible oyster distributions for each analyte. Probability distributions of the means and the upper 95th percentile were obtained for these oyster data ensembles and are presented in Chapter 6. A comparison of simulation results with results predicted by the EqP model clearly demonstrates a divergence of outcomes as the molecular weight of the compounds increases. In the case of benzo{a}pyrene, for example, the mean and upper 95th percentile estimates from the EqP model are over an order of magnitude larger than those obtained by Monte Carlo simulations. Since the high molecular weight P AHs pose the greatest threat to the human consumer, use of the EqP model may greatly overestimate the health risks from seafood consumption.

Rights

All rights reserved. All rights reserved. Copyright is held by the author.

Download

Share

COinS