Episodic stream acidification occurs as storm events temporarily reduce acid neutralizing capacity and pH. Stream acidification is suspected to have damaging effects on the health of aquatic ecosystems and biota and is dependent on various watershed characteristics such as drainage area, elevation, slope, and surficial geology. Here, a stochastic modeling approach is applied to continuous pH data for multiple stream monitoring sites within the Great Smoky Mountains National Park in order to characterize episodic acidification responses during stormflows for different streams. The approach summarizes voluminous pH data recorded by water quality sondes at 15-minute intervals into concentration-duration-frequency relationships. Unique to this study is the ability to characterize the episodic acidification response to watershed attributes without using baseflow or single-point stormflow measurements. A slope metric of mean pH event duration, a measure of episodic acidification response was determined to correlate with basin area and elevation. In contrast, baseflow studies have shown elevation to be the main driver of chronic acidification. It appears that during stormflows transport and flushing of stored anions and cations govern the response of streams included in this study.

In order to accurately predict the stability of riverbanks, model input parameters must be reliable bank failure estimators. Currently, bank stability models require two input parameters to predict bank erosion: critical erosion shear stress and erodibility coefficient. The investigation’s purpose was to compare two erosion estimation methods and improve the bank stability models for cohesive soil commonly found on the banks. To accomplish the objective, critical shear stresses and erodibility coefficients obtained using the in situ submerged jet test device (SJT) were measured against results from the closed-loop laboratory flume method for 12 cohesive bank sites. Additionally, SJT critical shear stress values were compared to values found via empirical relationships found in literature that incorporate plasticity index, median particle diameter, percent siltclay or percent clay content to compute critical shear stress. Particle size analysis and Atterberg limit determinations were run classify the sediment type collected. The critical shear stress values obtained ranged from 0.09 to 5.84 Pa and SJT erodibility coefficients varied from 0.37 to 10.07 cm3/N·s. From flume observations, cohesive soil erosion was influenced by interparticle forces and occurred in aggregate pieces and particle-by-particle. A few critical shear stress values appeared to be unreliable considering the critical shear stress threshold of 1.83 Pa found using the laboratory flume analysis and the limited erosion witnessed. Study results also indicated that sediment properties did not correlate directly with the SJT critical shear stress values or with each other. Flume observations and variations among experimental results suggest other influential factors exist besides critical shear stress and the erodibility coefficient when quantifying the cohesive sediment erosivity. When empirical results were lower than the flume’s critical shear stress threshold, it was possible the mechanical soil property could not be transferred to the soil types tested or estimates incorrectly assumed zero physical and chemical influences. Because of its complexities, traditional experimental design may not reliably measure cohesive soil erosion. Only through the continued collaboration of various field and advanced degree professionals and the detailed, high-quality documentation of as many influential parameters as possible per project can the goal of estimating cohesive sediment erosion be accomplished.

This study attempts to develop a relationship with the hillslope sediment yield (estimated from a computer model) and the deposited sediment particle size characteristics within stream channels. By using specific hydrological parameters within a watershed, a calibrated Annualized Agricultural Non-Point Source (AnnAGNPS) pollutant loading model was created for four different sub-watersheds in the mountainous New River Basin of eastern Tennessee. The AnnAGNPS pollutant loading model predicted daily runoff and sediment yield reasonably well, but it poorly predicted daily peak flow rate for most sub-watersheds analyzed in the New River Basin. Overall, the AnnAGNPS pollutant loading model provided satisfactory results in a mountainous, nonagricultural landscape with a limited amount of climatic data available. The average annual hillslope sediment yield, in terms of clays, silts, and sands, was calculated with the AnnAGNPS model for years 2006 and 2007, to compare with sediment deposition characteristics in the streams. The fine particle size characteristics collected at specific bed deposition points were suspected to have a strong correlation with predicted sediment yield output from a calibrated AnnAGNPS pollutant loading model. The sites of the captured sediment were at locations just downstream of specific land use disturbances such as dirt roads, surface mining, and forest logging, all of which can be detrimental to the health of a stream environment and habitat if disturbances are not properly managed. In this study, the sediment collected at the channel bed deposition points represented the distribution of different material sizes that have recently moved within the stream during large discharge events. This investigation concluded that the certain measurements of the clays, silts, sands, and gravel material found in downstream sediment depositional points had a variety of significant relationships (p-value < 0.05) with the clays, silts, sands, and total sediment yield occurring on the watershed hillslopes. Overall, there are a limited amount of studies that analyze these collections of fine sediment deposited in areas of the stream that have interrupted velocity forces due to channel shape, objects, or formations. This study showed that the use of the AnnAGNPS pollutant loading model and the analyzation of specific fine sediment at depositional points in the stream, proper watershed management of a rural mountainous region can be better established.

Groundwater-surface water interactions have been shown to be important to flow generation and stream chemistry in upland catchment environments. These areas, however, are often difficult to access making the implementation of standard hydrological surface and subsurface monitoring equipment and characterization procedures impractical, arduous and in many cases impossible due to the nature of the terrain and also regulatory guidelines for protected areas. By collecting surface water samples at distinct water contribution sites to a headwater stream, areas of groundwater influence were inferred and a hydrochemical conceptual model of a small basin was created. The objectives of this research were to 1) understand the groundwater chemistry influences to an upland stream in the Great Smoky Mountains National Park (GSMNP) using limited surface water data 2) determine if the use of multivariate statistics could help delineate water interaction “types” within the study basin, 3) create a conceptual model to define the chemical interactions in the stream using a comparison of data. The objectives were met by the analyses of field data collected within Ramsey Prong, a remote forested, high elevation stream in the Middle Prong Little Pigeon River Watershed chosen for the study site. Eight sampling sites were selected at hydrologic and hydrochemically significant points in the basin. Three data collection trips were performed in April, July, and August of 2007. Water sampled for analysis of cations, anions, and trace metals was collected and flow measurements were recorded on each trip at each site. Multivariate analyses were conducted on the collected data to detect correlations between parameters that might indicate similar chemistries or water interaction “types” where high correlations were displayed. Three water types: 1) surface water; 2) spring water; and 3) a top of catchment mixture of spring and highly acidic deposition and drainage water were delineated. Spring 1, located at the bottom of the study area, was designated as the first water type and displayed high concentrations of Si, Na, ANC, and pH. The source of this water was affirmed by groundw32ater characteristics caused by the sandstone subsurface environment. The second water type, consisting of the two highest elevation sample sites, displayed characteristics of acid deposition and acid-rock or acid induced leaching including, low pH and ANC and high levels of SO4, Mn, Fe, and Al ions. Increased levels of Si and Na also suggested groundwater interaction further up the sampled tributary. Water designated as the third water type consisted of the remaining in-stream samples which demonstrated a trend of general dilution in most atmospherically input ions and a concentrating of geochemical parameters. Large areas of focused recharge signatures were not detected in the study area leading to the assumption of primarily diffuse recharge throughout the stream. Two tributaries sampled on the August collection date displayed groundwater chemistries with a different signature than that of the stream and showed indications of water quality buffering. The analysis demonstrated the possibility for influential stream buffering in the GSMNP by groundwater and groundwater sourced inputs and also the importance of groundwater in this system.

The primary purpose of this research effort was to examine the dry removal of SO2 and NO from a flue gas stream by injecting two sodium additives into a pilot bag house system. The additives tested were NaHCO3 and Na2CO3 dusts with mass mean diameters ranging from approximately 30 to 200 microns. The Na2CO3 was obtained by decomposing the NaHCO3 (with heat) prior to testing. The bag house temperature was maintained at either 250 or 300 degrees F. It was demonstrated that 70% SO2 removal can be attained with NaHCO3 powders that have mass mean diameters of 32 and 52 microns at stoichiometric ratios of 0.8 and 1.3, respectively. The rate of mass transfer limited the desulfurization capacity of NaHCO3 powders with mass mean diameters greater than 50 microns. The rate of chemical reaction limited the SO2 removal capability of the smallest NaHCO3 additive tested (mass mean diameter = 32 microns). Decomposition of NaHCO3 to Na2CO3 in bulk before injection yielded poorer SO2 removal. It was also demonstrated that 7 to 36% of the NO was removed simultaneously with SO2 by the NaHCO3 additives with mass mean diameters smaller than 120 microns. This removal was inversely dependent on the system temperature. No appreciable NO removal was observed with Na2CO3 injection.

Pharmaceuticals and Personal Care Products (PCPPs) are an emerging class of water pollutants, which have the potential for human health and adverse environmental impacts. This thesis will seek to answer key questions related to biological treatment, adsorption processes, the viability of advanced treatment processes, and business considerations to remove PCPPs from water. The hypotheses are: How does the concentration of these micropollutants affect their ability to be treated? How are carbon adsorption and biological treatment affected by competition of other substrates? To what extent will current technology allow for their safe removal from discharge sources? Addressing these questions will be accomplished with Monte Carlo simulations and kinetics modeling. To successfully remove these micropollutants, biological, chemical adsorption, and other novel treatment methods were explored to ascertain if these methods could be utilized to remove these substance from water and wastewater. It was found that currently available treatment methods (biological and chemical adsorption) are highly dependent upon the competitive effects that occur as a result of the other constituents within wastewater. Therefore, advances processes like wet-air oxidation, Reverse Osmosis, and Plasma Arc Waste Disposal must be considered to adequately remove these compounds.

Near-surface contaminant isolation facilities are routinely used for the disposal of hazardous and radioactive waste. The objective of these facilities is to maintain the long-term isolation of the identified contaminants as well as to mitigate their associated hazards. Societys experience with modern isolation techniques is beginning to show that these facilities and associated management techniques do not always perform as expected. This research investigated the potential failure of contaminant isolation facilities. Failure, in the context of this research refers to a contaminant isolation facilitys inability to maintain contaminant isolation. Errors were viewed as potential precursors to future failure as they are an indication that actions did not go according to the plan or that the plan itself was inadequate to achieve the objective. This research used a multiple case study design to further investigate what barriers and controls are used for long-term contaminant isolation and how these controls are currently performing. Seven case study reports were developed as part of this research. The sites investigated included Anaconda, Burrell, Canonsburg, Love Canal, Maxey Flats, Rocky Mountain Arsenal, and Spring Valley. Each of these cases involved sites where persistent contaminants were isolated on-site in the shallow subsurface. A cross-case failure analysis was then performed. Fault trees were developed and are presented in this research to illustrate potential failure pathways for each of the contaminant isolation facility controls. Individual controls were found to contain many single point errors. These controls can be described as non-robust in that one event could potentially lead to a control error. System failure was found to require the breach of an engineered barrier. For engineered barrier error to occur a prior institutional control error must occur. Institutional control error alone does not appear to lead to system failure. Institutional control error, however, could be viewed as precursors to potential system failure. The lack of information management, stakeholder awareness and adequate resources were found to be key initiating events potentially leading to system failure. These initiating events could lead to latent institutional control error, which are not fully realized until subsequent events occur.

Secondary materials such as bottom ash, slag, concrete debris, and other waste materials are being considered for use as substitutes for natural aggregates in highway applications due to suitable engineering and economic properties. This research investigated the effect of freezing and thawing on flow mechanisms and constituent leaching from recycled concrete during a scenario of infiltration. Laboratory flow column tracer and leaching studies coupled with freeze/thaw aging were performed on laboratory formulated concrete. Three material moisture contents including the optimum moisture content of the packed material and three levels of freeze/thaw aging were examined to simulate different extents of freeze/thaw exposure. Variation in flow through the columns was observed between the different levels of freeze/thaw cycling and moisture contents. The columns receiving the maximum number of freeze/thaw cycles demonstrated a lower release of calcium, sodium, potassium, chloride, and sulfate, especially at low LS ratios, which may be due to consolidation of the material during freezing and thawing action. Overall, no significant effect of freeze/thaw cycling was observed in the leaching of trace metals. Results were compared to columns subjected to more rapid freeze/thaw cycling. Additionally, a limited number of intermittent flow columns with F/T aging were run on both laboratory formulated concrete material and actual recycled concrete (construction debris) to simulate field-like conditions.

Landfill covers systems are designed to isolate contaminants from the environment for hundreds to thousands of years. In the long term, however, the landfill covers are susceptible to failure caused by physical, chemical and biological effects. Plant intrusion is a major concern of landfill sustainability; the roots of plants have been found to increase the soil hydraulic conductivity by two orders of magnitude. The objective of this thesis was to evaluate the effect of plant intrusion and the resulting increase in the saturated hydraulic conductivity on the water balance of landfill cover systems. The finite element software model HYDRUS-1D was used to simulate water flux and root water uptake processes. Plant growth effects on the Burrell, Pennsylvania, uranium mill tailings site were evaluated as a case study. Additionally, a comparison of the water balance and effects of plant growth for different landfill cover designs (RCRA Subtitle C, RCRA Subtitle D and evapotranspiration covers) was performed. For each of the cases, different precipitation events were used to simulate arid and humid conditions. The plant roots effectively reduced the water content in the landfill cover systems, and it was concluded that vegetation should not be eliminated once established.

The inherent heterogeneity of subsurface porous media, as well as the occurrence of highly non-linear rate-limiting mass transfer processes, results in significant challenges to accurate and efficient modeling of contaminant flow and transport. This dissertation provides comprehensive Web-based modeling tools and advanced numerical methods for students and researchers to better investigate fluid flow and mass transfer processes in natural and model systems under water-saturated conditions. A Web-based virtual mass transfer processes laboratory (MTVLab) was developed for students and researchers to access and study state of the art understanding of mass transfer mechanisms at the particle scale. Meanwhile, MTVLab system architecture provides a proof-of-principle framework from which to develop more sophisticated Web-based models that can employ computationally efficient, high-level computer programs. MTVLab is available at http://www.vanderbilt.edu/mtvlab. Lattice Boltzmann methods (LBM) were used to study fluid flow in two-dimensional randomly generated porous media. An innovative method was developed to construct permeability cumulative distribution functions through the combination of LBM and first order reliability method (FORM). LBM FORM was found to be approximately 13 to 120 times more efficient than traditional Monte Carlo-based simulations while maintaining similar orders of accuracy. A novel least squares finite element lattice Boltzmann method (LSFE-LBM) was developed, extending LBM to unstructured meshes. LSFE-LBM is able to more efficiently simulate fluid flow and solute transport in domains that contain complex or irregular geometric boundaries. LSFE-LBM provided the foundation for the numerical modeling efforts to elucidate the relative contributions of transport-related and sorption/desorption - related nonequilibrium factors on mass transfer processes in a whole class of porous media exemplified by randomly generated porous media. Applications of LSFE-LBM to simulate phenanthrene transport in porous media represent an initial effort to bridge comprehensive sorption/desorption mechanistic studies with porescale modeling, the results of which help advance our understanding of the effects of soil organic matter and soil structure configurations on fate and transport of organic chemicals in subsurface systems.