Time-Lapse Geophysics and Characterization of an Infiltration Experiment in Unsaturated Wisconsin Alluvial Sediments
File(s)
Date
2021-08Author
Swager, Katherine
Advisor(s)
Fratta, Dante
Metadata
Show full item recordAbstract
This thesis discusses the use of field geophysics, laboratory testing, and numerical
modeling to characterize the infiltration mechanisms of a plume in the vadose zone. The study
of fluid flow and transport in the vadose zone is critical to understand the fate of agriculturally
derived contaminants such as nitrate and pathogens as they move from the surface to the
vadose zone, and then into the groundwater.
Agricultural contamination is common in private wells in rural Wisconsin, especially in
aquifers with a high effective porosity, such as in the Lower Wisconsin River Valley where this
research was conducted. Three infiltration field experiments were completed on the noncultivated
edge of an agricultural field in Spring Green, WI where sediments consist of sand and
gravel from glacial outwash. A constant head infiltrometer was used to permeate a bromide
tracer solution into the soil surface. Time-lapse ground penetrating radar (GPR) and electrical
resistivity tomography (ERT) surveys were collected on a grid above the infiltration.
Subsequently, soil samples were collected and characterized by water content, grain size,
permeability, irreducible water saturation, and pore water tracer concentration.
Initial results from the electrical resistivity surveys indicated the soil may be a two-layer
system, with a transition to a coarser grain size at 0.75 m where the infiltration fluid ponded.
The ground penetrating radar showed reflectors at approximately 1 and 2 m depth. The
infiltration was modeled in TOUGH2. A 3-D, two-layer model with a fine sand overlying a coarse
sand was constructed using laboratory and literature data. The model results showed that the
water would break through to the lower layer in the timeline shown in the ERT data. However,
the two-layer model could not explain the volumetric water content results, and the ERT data
10 was not supported by the chemistry or water content. Grain size distribution analysis showed
that there may be multiple textural soil layers. A multi-layer model was formulated to
represent the grain size distributions, and the results were more representative of the
laboratory water content and bromide concentrations than the two-layer model.
Further work including forward modeling should be done to understand the relation of the
geophysical signals to the hydrologic conditions and soil stratigraphy. Further laboratory soil
testing could increase the understanding of the site’s heterogeneity. The presence of textural
layering in soils may have implications for the timeline of contaminant transport into the water
table.