Mary P. Anderson

Using models to simulate the movement of contaminants through groundwater flow systems

Prediction of the movement of contaminants in groundwater systems through the use of models has been given increased emphasis in recent years because of the growing trend toward subsurface disposal of wastes. Prediction is especially critical when nuclear wastes are involved. Contaminant transport models which include the effects of dispersion have been applied to several field situations. However, factors that limit the routine use of these models include the difficulty of determining the field coefficient of dispersion and numerical difficulties encountered when solving the dispersion equation. Regional size models which neglect the effects of dispersion have had limited success because of the scarcity and poor quality of field data. Another difficulty in the development of contaminant transport models is the current lack of knowledge regarding the quantification of chemical reaction terms. This review examines the formulation of contaminant transport models, application to field problems, difficulties ...

Hydrogeologic facies models to delineate large-scale spatial trends in glacial and glaciofluvial sediments

Recent interest in contaminant transport in ground water has led hydrogeologists to the conclusion that predicting the movement of solutes requires information on the distribution of spatial trends and heterogeneities in porous media. Description of spatial trends has long been of interest to sedimentologists who have produced a large body of geologic information on the subject. In this paper, facies models are used to construct conceptual models of hydrogeologic facies for glacialmeltwater-stream sediment and till. These hydrogeologic facies models, which delineate large-scale trends in heterogeneity, are appropriate for use in designing hydrogeologic field tests and for estimating input to regional ground-water flow and transport models. This paper treats each facies as a homogeneous, anisotropic hydrogeologic unit. The models presented herein conceptualize the hydrogeologic relationships among facies and illustrate one method of converting the apparent chaos in nature into an orderly system that can be tested scientifically and modeled mathematically. The principles used to create conceptual models of hydrogeologic facies for the types of sediment considered in this paper can be extended to other sedimentary environments. It should be recognized, however, that such models do not address the small-scale heterogeneity present within individual facies. Additional basic research is required to measure hydraulic conductivity variation within representative hydrogeologic facies and to develop statistical descriptions to represent the variations. Such detailed descriptions of hydraulic conductivity may be necessary to describe ground-water flow at a local scale for analysis of contaminant transport.

Identifying spatial variability of groundwater discharge in a wetland stream using a distributed temperature sensor

[1] Discrete zones of groundwater discharge in a stream within a peat-dominated wetland were identified on the basis of variations in streambed temperature using a distributed temperature sensor (DTS). The DTS gives measurements of the spatial (±1 m) and temporal (15 min) variation of streambed temperature over a much larger reach of stream (>800 m) than previous methods. Isolated temperature anomalies observed along the stream correspond to focused groundwater discharge zones likely caused by soil pipes within the peat. The DTS also recorded variations in the number of temperature anomalies, where higher numbers correlated well with a gaining reach identified by stream gauging. Focused zones of groundwater discharge showed essentially no change in position over successive measurement periods. Results suggest DTS measurements will complement other techniques (e.g., seepage meters and stream gauging) and help further improve our understanding of groundwater-surface water dynamics in wetland streams.

Simulation of Preferential Flow in Three-Dimensional, Heterogeneous Conductivity Fields with Realistic Internal Architecture

Subsurface flow is primarily controlled by the distribution of hydrogeological properties. Spatially correlated random fields, currently the primary means of describing heterogeneity in these properties, underutilizes the existing knowledge of geological systems. A new computer code, the Braided Channel Simulator (BCS-3D), is used to geometrically simulate the spatial heterogeneity of hydrogeological properties using representations of surface topography to estimate the three-dimensional arrangement of subsurface units. Four methods of assigning conductivity were used to generate conductivity fields as a basis for saturated flow simulation. Variograms measured for each of the conductivity fields give no indication of underlying discrete structure. However, particle-tracking simulations showed that flow occurred along paths of preferentially high hydraulic conductivity. Results indicate that in systems where the contrast in conductivity is sufficiently great, the location and magnitude of flow is constrained by discrete internal structure. Additionally, geometrical simulation techniques such as BCS-3D can be used to produce property fields that embody these discrete geological structures.

Groundwater Inflow Measurements in Wetland Systems

Our current understanding of wetlands is insufficient to assess the effects of past and future wetland loss. While knowledge of wetland hydrology is crucial, groundwater flows are often neglected or uncertain. In this paper, groundwater inflows were estimated in wetlands in southwestern Wisconsin using traditional Darcys law calculations and three independent methods that included (1) stable isotope mass balances, (2) temperature profile modeling, and (3) numerical water balance modeling techniques. Inflows calculated using Darcys law were lower than inflows estimated using the other approaches and ranged from 0.02 to 0.3 cm/d. Estimates obtained using the other methods generally were higher (0.1 to 1.1 cm/d) and showed similar spatial trends. An areal map of groundwater flux generated by the water balance model demonstrated that areas of both recharge and discharge exist in what is considered a regional discharge area. While each method has strengths and weaknesses, the use of more than one method can reduce uncertainty in the estimates.

Groundwater's dynamic role in regulating acidity and chemistry in a precipitation-dominated lake

Abstract The role of groundwater in regulating the hydrologic and chemical regime ofan “acid-susceptible” lake situated in a sandy silicate aquifer in northern Wisconsin was determined by direct groundwater measurements using a network of 68 piezometers. Groundwater inputs were compared with precipitation and dryfall components of the hydrologic and chemical budgets for the lake. Groundwater contributed only 10% of the water to Crystal Lake in 1982 but was the major source of alkalinity, calcium, magnesium, sodium, potassium, silicon, and chloride. Precipitation contributed 90% of the water and was the major source of hydrogen and sulfate ions. Continuing analyses suggest that groundwater consistently supplies the bulk of the inoorganic constituents to the lake on annual basis. The flux of groundwater and dissolved solids to the lake is highly seasonal, with large pulses of groundwater inflow occuring after spring snowmelt and in the fall. For certain nonconservative elements such as silicon, the periodic influx of chemically concentrated groundwater can trigger significant changes in the chemical and biological balance in the lake. Despite the fact that the aquifer contains no carbonate minerals and the area experiences acid precipitation (pH = 4.6), the alkalinity of the small amount of groundwater entering Crystal Lake is more than enough to neutralize all of the acidity contributed by precipitation. In-lake sulfate reduction may generate an equal amount of alkalinity. This buffering of pH helps to regulate aluminum concentrations at low levels, thus minimizing potential aluminum toxicity in the lake. Most lakes in the region probably receive even larger amounts of alkalinity-rich groundwater than Crystal Lake, because most lie further downgradient in the regional flow system. This may help to explain why most lakes in this “acid-susceptible” area are no more acidic today than they were 50 years ago.

Sedimentology and hydrogeology of two braided stream deposits

Abstract Two approaches were used to quantify the spatial distribution of hydrofacies in braided stream deposits. One approach involved mapping a 50 by 60 by 3.3 m section of a proximal braided stream deposit. In a second study, we generated a 400 by 400 by 2.6 m section of a medial braided stream deposit using a computer model. In both cases we produced three-dimensional images showing connected hydrofacies with high permeabilities that form preferential flow paths. This information was input to a groundwater flow model and flow paths were analyzed by following the transport of imaginary particles. In both systems, particles that were uniformly distributed at the up-gradient end of the model clustered along preferential flow paths during transport, showing that connection among high-permeability facies is a critical factor in hydrogeological investigations involving assessment of contaminant movement and remediation.

The role of the postaudit in model validation

Abstract Most researchers agree that validation is a demonstration that a model is capable of making accurate predictions at a site-specific field setting. A successful demonstration of validation requires completion of a series of steps that form a modeling protocol. These steps include model design and calibration, and verification of the governing equation, the computer code, and the model itself. The strictest form of validation is to demonstrate that the model can accurately predict the future. This type of validation test has been called a postaudit. Results of five postaudits suggest that it will be difficult and probably impossible to validate groundwater models by means of a postaudit because it is impossible to characterize the field setting in sufficient detail. Attention should instead be focused on good modeling protocol including providing a complete description of model design, a thorough assessment of model calibration, and an uncertainty analysis.

Long- and short-term transience in a groundwater/lake system in Wisconsin, USA

Abstract A 10 year record of water level fluctuations in a groundwater/lake system in northern Wisconsin shows that system dynamics are strongly influenced by seasonal transient effects as well as transience over the 10 year period of record. These data were collected for the Long Term Ecological Research (LTER) program at the North Temperate Lakes site in northern Wisconsin, USA. The record included a period of relatively high lake levels from 1981 to 1985, followed by declining levels in 1986–1987 and low levels in 1988–1990. Short-term transient effects in the form of seasonal groundwater mounds consistently occurred on all sides of the lake during 1981–1985 when regional water levels were high and during 1986–1988 when groundwater levels were high relative to the declining lake level. Mounds did not form after 1988. These observations point to the importance of a long-term record in assessing the significance of short-term effects. Short-term transience affects the groundwater component of the lake budget because the mounds induce groundwater to flow toward the lake; when the mounds are not present, water flows away from the lake. Shifts in the groundwater regime will affect the lakes chemical budget in the long term. However, the trends in the chemical budget that occurred within our 10 year record, an interval shorter than the hydraulic residence time of the lake (12.7 year), are influenced more by changes in precipitation inputs than groundwater inputs.

Importance of unsaturated zone flow for simulating recharge in a humid climate.

Transient recharge to the water table is often not well understood or quantified. Two approaches for simulating transient recharge in a ground water flow model were investigated using the Trout Lake watershed in north-central Wisconsin: (1) a traditional approach of adding recharge directly to the water table and (2) routing the same volume of water through an unsaturated zone column to the water table. Areas with thin (less than 1 m) unsaturated zones showed little difference in timing of recharge between the two approaches; when water was routed through the unsaturated zone, however, less recharge was delivered to the water table and more discharge occurred to the surface because recharge direction and magnitude changed when the water table rose to the land surface. Areas with a thick (15 to 26 m) unsaturated zone were characterized by multimonth lags between infiltration and recharge, and, in some cases, wetting fronts from precipitation events during the fall overtook and mixed with infiltration from the previous spring snowmelt. Thus, in thicker unsaturated zones, the volume of water infiltrated was properly simulated using the traditional approach, but the timing was different from simulations that included unsaturated zone flow. Routing of rejected recharge and ground water discharge at land surface to surface water features also provided a better simulation of the observed flow regime in a stream at the basin outlet. These results demonstrate that consideration of flow through the unsaturated zone may be important when simulating transient ground water flow in humid climates with shallow water tables.