Tracer Migration in a Radially Divergent Flow Field: Longitudinal Dispersivity and Anionic Tracer Retardation

Abstract

Hydrodynamic dispersion, the combined effects of chemical diffusion and differences in solute path length and flow velocity, is an important factor controlling contaminant migration in the subsurface environment. However, few comprehensive three‐dimensional datasets exist for critically evaluating the impact of travel distance and site heterogeneity on solute dispersion, and the conservative nature of several commonly used groundwater tracers is still in question. Therefore, we conducted a series of field‐scale experiments using tritiated water (3H1HO), bromide (Br−), and two fluorobenzoates (2,4 Di‐FBA, 2,6 Di‐FBA) as tracers in the water‐table aquifer on the USDOE's Savannah River Site (SRS), located on the upper Atlantic Coastal Plain. For each experiment, tracer‐free groundwater was injected for approximately 24 h (56.7 L min−1) to establish a steady‐state forced radial gradient before the introduction of a tracer pulse. After the tracer pulse, which lasted from 256 to 560 min, the forced gradient was maintained throughout the experiment using nonlabeled groundwater. Tracer migration was monitored using six multilevel monitoring wells, radially spaced at approximate distances of 2.0, 3.0, and 4.5 m from the central injection well. Each sampling well was further divided into three discrete sampling depths that were pumped continuously (∼0.1 L min−1) throughout the course of the experiments. Longitudinal dispersivity (αL) and travel times for 3H1HO breakthrough were estimated by fitting the field data to analytical approximations of the advection–dispersion equation (ADE) for uniform and radial flow conditions. Dispersivity varied greatly between wells located at similar transport distances and even between zones within a given well, which we attributed to variability in the hydraulic conductivity at the study site. The radial flow equation generally described tritium breakthrough better than the uniform flow solution, as indicated by the coefficient of determination, r2, yielding lower αL while accounting for breakthrough tailing inherent to radial flow conditions. Complex multiple‐peak breakthrough patterns were observed within certain sampling zones, indicative of multiple major flow paths and the superposition of resulting breakthrough curves. A strong correlation was found between αL and arrival times observed from one experiment to the next, indicative of the general reproducibility of the tracer results. Temporal moment analysis was used to evaluate tracer migration rate as an indicator of variations in hydraulic conductivity and flow velocity, as well as mass recovery and retardation for the ionic solutes compared with tritiated water. Retardation factors for Br− ranged from 0.99 to 1.67 with no clear trend with respect to transport distance; however, Br− mass recovery decreased with distance, suggesting that the retardation values are biased in terms of early arrival because of limited detection and an insufficient monitoring duration. Anion retardation was attributed to sorption by iron oxides. Simila...

Spatial and Temporal Variability in Colloid Dispersion as a Function of Groundwater Injection Rate within Atlantic Coastal Plain Sediments

“…As might be expected for a system displaying amphoteric surface charge, minimal colloid generation occurred when the pH (∼5.0–5.5) and solution composition were similar to the low ionic strength groundwater typical of the water‐table aquifer on the SRS. Additionally, the duration of our study was insufficient to alter the groundwater chemistry in all but the closest monitoring wells, as confirmed by subsequent groundwater tracer studies (Seaman et al, 2007). Therefore, colloid mobilization here likely resulted from a combination of formation disturbance during well installation and the altered hydraulic gradient associated with injection and groundwater sampling.…”

“…Therefore, our objectives were (i) to evaluate the impact of injection rate on colloid generation as a function of depth within the formation and distance from the injection well and (ii) to evaluate the impact of colloid mobilization on water quality and formation permeability in the vicinity of the injection well. We used information gained during the initial test to plan and design subsequent long‐term injection experiments to evaluate the impact of solution chemistry on colloid mobilization, as well as for several extensive groundwater tracer experiments (Seaman et al, 2007).…”

When “Conservative” Anionic Tracers Aren't

“… 3 H 2 O is more conservative than Br − in anion‐sorbing sediments (Seaman et al 1995; Korom 2000). Br − adsorption is nonlinear (i.e., concentration‐ dependent), resulting in greater tailing (i.e., dispersion) than apparent for a “conservative” tracer (Seaman et al 1995; Seaman et al 1996; Korom 2000; Seaman et al 2007). Korom (2000), in a column study using an undisturbed sand with Br − concentrations from <0.1 to 700 mg/L, found the sorption was best fit by a second‐order polynomial; however, a Freundlich isotherm also described the adsorption within a 95% confidence interval on the regression. Br − sorption increases with increasing concentrations (Korom, 2000), as suggested by Figure 1B with pH < PZC.…”

“…The variances of the natural logarithms of the hydraulic conductivity fields were varied over an order of magnitude and two operationally defined parameters were considered: the amount of Br − injected and the lowest tracer concentration value used in the dispersivity computations as a result of analytical quantification limits for the tracer or because of measureable background concentrations of the tracer (Korom and Dodak 2009). Using two FBA tracers, Br − , and 3 H 2 O in a series of field‐scale forced‐gradient tracer tests, Seaman et al (2007) found that the assumption of conservative behavior for the anionic tracers generally yielded larger longitudinal dispersivity values than did the experiments using 3 H 2 O. They also concluded that advection‐dispersion models reasonably “appeared” to describe the non‐conservative anion tracer transport without accounting for anion sorption, indicating that geochemical interactions with the aquifer sediment may be interpreted incorrectly in terms of physical transport processes (i.e., flow velocity, dispersivity, multiple flow domains, etc.…”

Markierungsversuche und Modellierung zur Bewertung der Gefährdung eines Trinkwasserbrunnens

“…Diese können durch die Überlagerung mehrerer aDM-Kurven simuliert werden (z. b. Seaman et al 2007). Für jede einzelne Kurve müssen die Werte für t 0 und D bestimmt werden.…”