In 1990, creosote contamination was discovered at a railroad tie yard in Oneida, Tennessee. A phytoremediation system that included over 1,200 hybrid poplar trees was installed between 1997 and 1998 for hydraulic control of the groundwater and enhancement of the natural biodegradation processes in the subsurface. Since then, Virginia Polytechnic Institute and State University has monitored eight polycyclic aromatic hydrocarbons (PAHs) in the soil and groundwater. They have found that concentrations of smaller and more volatile PAHs have decreased over the years as the DNAPL contamination has become more enriched with the larger PAHs. This thesis focuses on the movement of naphthalene through the subsurface because it comprises the majority of the creosote and evidence for its remediation exists. Of the many mechanisms within the phytoremediation system that serve to remediate contaminated groundwater and soil, the most important are rhizosphere bioremediation and plant uptake. However, another mechanism, direct volatilization through the soil, was thought to have significant remediation capabilities at this site. Because naphthalene is a highly volatile PAH, it was hypothesized that naphthalene is volatilizing directly through the soil to the atmosphere and that the rate of volatilization may be enhanced by the presence of the phytoremediation system. The goals of this research are to measure the amount of naphthalene that volatilizes from the subsurface and determine the factors that significantly influence this direct volatilization. A flux chamber was designed and constructed to measure naphthalene fluxes from the soil. Factors that influence direct volatilization include the groundwater level, soil moisture, precipitation, pressure changes, temperature and humidity, the most important of which we found to be the groundwater level through its influence on naphthalene concentrations in the groundwater. We found that the presence of the trees significantly affects groundwater levels. As trees transpire and lower the groundwater table, concentrations in the uppermost portion of the groundwater increase, and under dry conditions, naphthalene fluxes from the soil are maximized. To complement the field measurements of direct volatilization, we also investigated rates of volatilization and biodegradation in the laboratory. Column experiments were conducted to determine the importance of direct volatilization on biodegradation in the vadose zone. We hypothesize that the combined mechanisms of contaminant transfer to the vadose zone, followed by rapid biodegradation, speeds up remediation in contrast to biodegradation that occurs only in the saturated zone under high groundwater conditions. Several columns using contaminated and uncontaminated soil from the site were constructed with a naphthalene source. Vertical naphthalene vapor concentration profiles were measured, and first-order biodegradation rates were determined. We found that biodegradation rates in the bacterially active columns were small initially, but that the biodegradation rates of the contaminated soil dramatically increased at day 60, while the biodegradation rates of the uncontaminated soil did not begin to increase until day 150. By the end of the experiment, both soil types had approximately the same biodegradation rate, signifying that soil that had previously been exposed to naphthalene degrades naphthalene more efficiently in the early stages than soil that has not been exposed, but that over time the non-exposed soil degrades naphthalene as efficiently as the pre-exposed soil. We determined that the combined mechanisms of diffusion and biodegradation in the unsaturated zone have significant remediation capabilities. Because long-term exposure risks are associated with inhaling indoor contaminant vapors, the Johnson and Ettinger vapor intrusion model was applied to the creosote-contaminated site, as outlined in Appendix C. This model takes into account soil, chemical, and building foundation characteristics to determine a dimensionless attenuation ratio, which is the ratio of contaminant vapor concentration in an enclosed space (i.e. basement) to the vapor concentration directly above the source. For a conservative case, the Johnson and Ettinger model without biodegradation was used. We found that if the land were developed, naphthalene vapor intrusion would not pose any health risks based on regulatory standards and levels at which health effects have been recorded. / Master of Science