Macropore flow and transport dynamics in partially saturated low permeability soils

Abstract

Near-surface sediments play an important role in governing the movement of water and contaminants from the land surface through the vadose zone to groundwater. Generally, low permeability surficial soils restrict water flow through the vadose zone and form a natural protective barrier to migration of surface applied contaminants. These types of fine-grained soils commonly contain macropores, such as fractures, animal burrows, and root holes, that have been identified as preferential flow pathways in the subsurface. Accordingly, macropores have the potential to influence groundwater recharge rates and compromise the protective capacity of surficial soils, particularly where the overburden is thin and aquifers are close to the surface. Partially saturated flow and transport in these environments is inherently complex and not well understood. The objective of this thesis was to examine preferential flow processes and the associated movement of contaminants in macroporous, low permeability soils. This was accomplished by conducting numerical and field experiments to investigate and describe the dynamics of macropore flow during episodic infiltration through the vadose zone and evaluate the corresponding influence of macropores on vertical water flow and contaminant transport.

Numerical simulations were conducted to identify the important physical factors controlling flow and transport behaviour in partially saturated, fractured soils. A three-dimensional discrete fracture model, HydroGeoSphere, was used to simulate infiltration into homogeneous soil blocks containing a single vertical rough-walled fracture. Relatively large rainfall events with return periods ranging from 5 to 100 years were used, since they are more likely to generate significant preferential flow. Initial results showed that flow system dynamics were considerably more sensitive to matrix properties, namely permeability and antecedent moisture content, than fracture properties. Capillary forces, combined with the larger water storage capacity in the soil matrix, resulted in significant fracture-matrix interaction which effectively limited preferential flow down the fracture. It is also believed that fracture-matrix interaction reduced the influence of fracture roughness and other related small-scale fracture properties. The results imply that aperture variability within individual fractures may be neglected when modeling water flow through unsaturated soils. Nevertheless, fracture flow was still an important process since the fracture carried the majority of the water flow and virtually all of the mass of a surface applied tracer to depth in the soil profile.

Model runs designed to assess transport variability under a variety of different physical settings, including a wider range of soil types, were also completed. Vertical contaminant fluxes were examined at several depths in the soil profile. The results showed that the presence of macropores (in the form of fractures) was more important than matrix permeability in controlling the rate of contaminant migration through soils. The depth of contaminant migration was strongly dependent on the antecedent moisture content and the presence of vertically connected fractures. Soil moisture content played a pivotal role in determining the onset and extent of preferential flow, with initially wet soils much more prone to macropore flow and deep contaminant migration. Simulations showed that surface applied tracers were able to reach the base of 2 m thick fractured soil profiles under wetter soil conditions (i.e., shallow water table). Likewise, long-duration, low-intensity rainfall events that caused the soil to wet up more resulted in proportionately more contaminant flux at depth. Fractured soils were particularly susceptible to rapid colloid movement with particle travel times to depths of 2 m on the order of minutes. The main implication is that the vulnerability of shallow groundwater is related more to vertical macropore continuity and moisture conditions in the soil profile, rather than traditional factors such as soil thickness and permeability.

Macropore flow and transport processes under field conditions were investigated using small-scale infiltration experiments at sites in Elora and Walkerton, Ontario. A series of equal-volume infiltration experiments were conducted at both sites using a tension infiltrometer (TI) to control the (negative) infiltration pressures and hence the potential for macropore flow. A simulated rainfall experiment was also conducted on a small plot at Walkerton for comparison with the TI tests. Brilliant Blue FCF dye and fluorescent microsphere tracers were applied in all tests as surrogates for dissolved and colloidal contaminant species, respectively. Upon completion of infiltration, excavations were completed to examine and photograph the dye-stained flow patterns, map soil and macropore features, and collect soil samples for analysis of microspheres. Cylindrical macropores, in the form of earthworm burrows, were the most prevalent macropore type at both sites. In the TI tests, there was a clear relationship between the vertical extent of infiltration and the maximum pressure head applied to the TI disc. Larger infiltration pressures resulted in increased infiltration rates, more spatial and temporal variability in soil water content, and increased depths of dye penetration, all of which were attributed to preferential flow along macropores. Preferential flow was limited to tests with applied pressure heads greater than -3 cm. Under the largest applied pressures (greater than -1.0 cm), dye staining was observed between 0.7 and 1.0 m depth, which is near the seasonal maximum water table depth at both field sites. The tension infiltrometer was also used to infiltrate dye along an exposed vertical soil face, thereby providing a rare opportunity to directly observe transient macropore flow processes. The resulting vertical flow velocities within the macropores were on the order of tens of meters per day, illustrating the potential for rapid subsurface flow in macropores, even under partially saturated conditions. The results suggest that significant flow occurred in partially saturated macropores and this was supported by simple calculations using recent liquid configuration models for describing flow in idealized macropores.

On all excavated sections, microspheres were preferentially retained (relative to the dye) in the top five centimeters of the soil profile. Below this zone, dye patterns correlated well with the presence of microspheres in the soil samples. There was evidence for increased retention of microspheres at lower water contents as well as a slightly greater extent of transport for smaller microspheres. In general, the microsphere and dye distributions were clearly dictated by vadose zone flow processes.

As in the numerical experiments, water storage in the soil matrix and related macropore-matrix interaction were important factors. Mass transfer of water through the macropore walls promoted flow initiation in the macropores near surface. Deeper in the soil, water drawn away from the macropores into the matrix significantly retarded the downward movement of water along the macropores. Imbibition of dye from the macropores into the matrix was repeatedly observed on excavated soil sections and during the transient dye test. Microspheres were also transported laterally into the soil matrix indicating that conceptual models for colloid transport in the vadose zone need to account for this mass transfer process.

Overall, the tension infiltrometer performed extremely well as a tool for controlling macropore flow under field conditions and, together with the dye and microsphere tracers, provided unique and valuable insights into small-scale flow and transport behavior. The field experiments raise concerns about the vulnerability of shallow groundwater in regions with thin, macroporous soils. Only a fraction of the visible macropores contributed to flow and transport at depths greater than 40 cm. However, with dye and microsphere transport observed to more than 1.0 m depth, rapid macropore flow velocities, and the sheer number of macropores present, there was clearly potential for significant flow and transport to depth via macropores. Under the right conditions, it is reasonable to speculate that macropores may represent a significant pathway for migration of surface applied contaminants to groundwater over the course of a single rainfall event.