Loading...
Thumbnail Image
Publication

Improved understanding of basic subsurface transport process interactions contributing to temporal variability in vapor intrusion: effect of infiltration and groundwater dynamics

Petri, Benjamin G.
Research Projects
Organizational Units
Journal Issue
Embargo Expires
2015-12-01
Abstract
Soil and groundwater contamination resulting from commercial and industrial activities at contaminated sites has long been known to pose threats to human health and the environment through multiple transport pathways. One of these is the vapor intrusion (VI) transport pathway, where vapor phase contaminants (especially volatile organic compounds) emanating from subsurface source zones or groundwater plumes migrate and intrude into buildings through foundation openings causing contamination of indoor air. To assess the risk of this transport pathway, a body of knowledge has been developed, emphasizing gathering of multiple lines-of-evidence, including sampling of indoor air, subslab soil gas, and other media in conjunction with screening models. However, these assessments are often challenged by uncertainty due to temporal variability in concentrations, especially within indoor air. Some of this temporal variability can result from the complex vadose environment where factors such as infiltration, water table dynamics and subsurface heterogeneity interact to affect the VI pathway. Subsurface vapor generation, transport and intrusion are ultimately controlled by a number of fundamental processes, including fluid flow in multiple phases (air, water, NAPL), mass transfer between phases, and simultaneous advection and diffusion within air and water phases. Much has already been reported in the literature on these individual processes themselves and their quantification. However, the complex interactions between these fundamental processes that occur in the presence of temporally varying factors such as an infiltration event or water table fluctuation have been only lightly explored in the literature, and are not incorporated in the conceptualization of the pathway vapor loading and development. Thus, the resulting "effective" behavior of the VI transport pathway in response to these temporal events is uncertain. With the objective of building a better conceptualization of the VI transport pathway that includes the dynamic response to these temporal factors, a study involving experiments and numerical models was initiated to generate data and gain an insight into these interactive processes, and their practical implications for the VI pathway. These experiments and models employed tightly controlled conditions and used only literature or independent measurements to parameterize models so that a phenomenological approach could be taken without relying on calibration or fitting parameters to match the data. Three manuscripts are contained within this thesis (Chapters 3, 4, and 5). These collectively explore the potential causes of spatial or temporal variability in the VI pathway through a sequence of experiments and numerical models, with implications for VI. The first manuscript explores the role soil moisture has on mass transfer from nonaqueous phase liquid (NAPL) sources entrapped in the vadose zone. These experiments determined that higher water contents within the immediate vicinity of a NAPL source could strongly limit mass transfer due to the higher diffusional resistance in the water versus gas phases. The implication for vapor intrusion is that fluctuations in water table elevations or infiltration that periodically exposes or occludes a NAPL source zone may cause strong changes in mass transfer from such a source. The second manuscript contains an experimental and numerical evaluation of how rain infiltration may dynamically interact with the VI pathway in a synthetic vapor intrusion setting. The experiment consisted of a large homogeneously packed sand tank configured with a vadose zone, water table and capillary fringe across which mass transfer occurs from a trichloroethylene contaminant plume. The top of the tank allows rain infiltration and atmospheric inflow while a synthetic building at the tank boundary draws are into the building. High temporal resolution concentration, soil moisture, air pressure and flow data are gathered throughout the experiment. Experimental data and modeling results highlight the importance of physical process interactions, where a spike in vapor intrusion occurs during infiltration events due to the displacement of soil gas by infiltrating water. After the spike, complex behavior that can include a washout or rebound effect can occur, and depends on the interaction between advection and diffusion. The third manuscript presents experiments conducted within the same experimental apparatus where the tank was subjected to a series of water table fluctuations. The purpose of specifically using a homogeneous, well-controlled, high temporal resolution experimental apparatus was to accurately record the pathway vapor loading due to water table fluctuation and test whether the models based on existing knowledge on fundamental process formulations can capture the observations without other complexities associated with subsurface heterogeneity. The observations suggest that water table fluctuations imparted a complex mass transfer behavior that defied simulation by a widely-used phenomenological modeling approach based on current knowledge. No arbitrary fitting process was used as is done in traditional model "validation" approaches. Instead, the data was used to gain insights on complex interactions to test the hypotheses that in this problem, the challenge remains to develop models that integrate the processes in a way to capture the complex interactions that contribute to temporal variability of vapor signals. These three manuscripts contain snapshots of the overall dataset, while the whole large tank experiment, including separate runs with both a homogeneous and heterogeneous packing are archived within appendix B (experimental methods and discussion) and appendix D (comprehensive data set) that will be of value for researchers in the future. It is hoped that these large data sets from the tank experiment will serve as a benchmark to improve conceptual and numerical models as well as serve for analysis of VI transport phenomena. The overall general conclusion from the experiments and models are that temporal vapor behavior can be driven by transient infiltration and groundwater dynamics, and improved conceptual and numerical models are needed to the full implications of these dynamic process interactions.
Associated Publications
Rights
Copyright of the original work is retained by the author.
Embedded videos