William W. Woessner

Measuring Groundwater–Stream Water Exchange: New Techniques for Installing Minipiezometers and Estimating Hydraulic Conductivity

Abstract Measurements of groundwater–stream water interactions are increasingly recognized as important to understanding the ecology of fishes and other organisms in stream and riparian ecosystems. However, standard measurement techniques are often feasible only at small spatial scales, in areas with easy access, or in systems with relatively fine substrata. We developed simple new techniques for installing minipiezometers and obtaining estimates of vertical hydraulic gradient, hydraulic conductivity, and specific discharge in gravel and cobble streambeds that allowed for large numbers of measurements to be obtained in remote locations. Our approach yielded values comparable to those obtained through more traditional methods. Consequently, these techniques may provide a labor cost-efficient way for detecting groundwater−stream water interaction patterns that are critical labor-attributes of stream and riparian systems at multiple scales.

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.

Ecohydrologic process modeling of mountain block groundwater recharge.

Regional mountain block recharge (MBR) is a key component of alluvial basin aquifer systems typical of the western United States. Yet neither water scientists nor resource managers have a commonly available and reasonably invoked quantitative method to constrain MBR rates. Recent advances in landscape-scale ecohydrologic process modeling offer the possibility that meteorological data and land surface physical and vegetative conditions can be used to generate estimates of MBR. A water balance was generated for a temperate 24,600-ha mountain watershed, elevation 1565 to 3207 m, using the ecosystem process model Biome-BGC (BioGeochemical Cycles) (Running and Hunt 1993). Input data included remotely sensed landscape information and climate data generated with the Mountain Climate Simulator (MT-CLIM) (Running et al. 1987). Estimated mean annual MBR flux into the crystalline bedrock terrain is 99,000 m(3) /d, or approximately 19% of annual precipitation for the 2003 water year. Controls on MBR predictions include evapotranspiration (radiation limited in wet years and moisture limited in dry years), soil properties, vegetative ecotones (significant at lower elevations), and snowmelt (dominant recharge process). The ecohydrologic model is also used to investigate how climatic and vegetative controls influence recharge dynamics within three elevation zones. The ecohydrologic model proves useful for investigating controls on recharge to mountain blocks as a function of climate and vegetation. Future efforts will need to investigate the uncertainty in the modeled water balance by incorporating an advanced understanding of mountain recharge processes, an ability to simulate those processes at varying scales, and independent approaches to calibrating MBR estimates.

The Impacts of Coal Strip Mining on the Hydrogeologic System of the Northern Great Plains: Case Study of Potential Impacts on the Northern Cheyenne Reservation

50% of the coal production of the U.S.A. will be obtained from the western coal fields by 1990. The majority of this coal will be produced from large-scale strip mining of the Tertiary Fort Union and Wasatch Formations of Wyoming, Montana and North Dakota. The rapid escalation of coal strip-mining activities in the Northern Great Plains, where groundwater is the principal source of domestic and agricultural supply, threatens to alter significantly the local and regional hydrologic regime. At the request of the Northern Cheyenne Tribe of southeastern Montana, a study of the potential impacts of strip mining on their water resources was designed and implemented. After a basic hydrogeologic study was conducted, a hypothetical mine study site was selected for evaluation. The pre-mining hydrologic system was defined from a monitoring well network, aquifer testing, water-quality sampling and stream gaging. Saturation extract and leachate analyses were conducted on cuttings of overburden from wells in the mine area and results were used to predict the concentration of total dissolved solids (TDS) in spoil-water discharge. A material-balance model was developed which described the quantity and quality of groundwater and recharge for zones of an unmined coal seam, spoil area, clinker and alluvium. A complementary material-balance model was also developed for a stream receiving post-mining discharge. The TDS of the spoil groundwater outflow was determined to be 4100 mg/l by averaging saturation extract data. Analyses of modeling results indicated that groundwater downgradient from the mined area could be increased in TDS by 300–2070 mg/l depending on the spoil recharge rate. The quality of the receiving stream with a mean annual flow of 14.2 m3/s could be increased in TDS by 1–26 mg/l. The site-specific mine-impact TDS changes were projected over an area of similar hydrogeology along 30 km of stream length in the Tongue River Valley. Strip mining of the entire minable area would have a major impact on the regional groundwater quality and a measurable impact on the quality of the receiving stream. Analysis of projected hydrologic properties of the post-mining system indicated that water-quality impacts will last for hundreds of years.

Chapter 9 – Model Calibration: Assessing Performance

Observations used to evaluate how well a model simulates a complex field situation are necessarily incomplete and error prone, which complicates the identification of a best (defensible) model from a family of reasonable models. Model evaluation, or calibration, uses both hard knowledge (field observations with simulated equivalent quantities) and soft knowledge (information not directly output by the model) to help identify a best model. Calibration uses history matching, which consists of changing model input (parameter values) and assessing the fit of the outputs (e.g., heads and fluxes) to field observations—a process that is similar whether done by manual trial-and-error methods or parameter estimation (inverse) software. Parameter estimation methods are powerful calibration tools that quantify the importance of individual observations, document a quantitative best model fit, and provide a quantitative foundation for forecast uncertainty analysis. Calibration of highly parameterized models is facilitated by advanced methods such as regularized inversion and singular value decomposition. Soft knowledge can be formally included in parameter estimation via Tikhonov regularization.

Basic Mathematics and the Computer Code

The governing equation for three-dimensional, transient groundwater flow in an anisotropic and heterogeneous porous medium is derived using conservation of mass and Darcys law. We briefly discuss fundamental concepts for analytical models, the analytic element method, and numerical models focusing on finite-difference and finite-element methods. We also discuss solution methods, convergence criteria, code selection and execution, graphical user interfaces, and the water budget computation.

Chapter 8 – Hyporheic Zones

Hyporheic zones include the saturated portions of streambeds, banks, and floodplain containing water that originates from a stream and returns to the channel. They are characterized by a mixture of local and regional groundwater and stream water, and typically vary in extent and duration. Hyporheic zones are habitat and refuge for various stages of aquatic organisms such as microbes, macroinvertebrates, and fish. They form ecotones that process solutes and carbon, support ecosystem metabolism, and influence stream biota and chemistry. Methods used to investigate and characterize hyporheic zones include instream piezometers, seepage meters, floodplain monitoring wells, steam gauging, tests to determine the hydrologic properties of saturated sediments, tracer studies, mapping of head distributions and flow directions, sampling of biota and geochemical constituents, modeling water exchange locations and rates, and defining geochemical cycling and heat exchange. Specific methods described are the installation of mini-piezometers, measuring the vertical hydraulic gradient (VHG), hydraulic conductivity and flux rates, estimating groundwater velocity, conducting and analyzing tracer tests, and analyzing floodplain monitoring well head and quality data sets.

Chapter 11 – The Modeling Report, Archive, and Review

The results of a modeling project are conveyed in a final report and included in an archive in sufficient detail to allow others to replicate and reproduce modeling results. The title of the report describes the modeling objective. The report includes an executive summary or abstract, an introduction, and sections on: hydrogeologic setting and conceptual model; numerical model; calibration; forecasting simulations and uncertainty analyses. The report summarizes the main findings and conclusions with respect to the modeling objective. The report, along with files containing input data and key output files and associated metadata, is included in the modeling archive. The archive should contain all the information necessary to replicate/reproduce and assess the modeling results. Modeling reports typically are peer reviewed.

Chapter 2 – Modeling Purpose and Conceptual Model

The modeling purpose underlies all subsequent steps in groundwater modeling. The purpose is defined by posing a question, or set of questions, which the model is designed to address. Most commonly the purpose is to forecast a response to changes in current conditions. The conceptual model, which provides a qualitative framework for designing a numerical model, is a descriptive representation of a groundwater system based on what is known about the modeled area. Development of a site conceptual model involves the synthesis of regional, site, and generic hydrogeologic and other relevant information. A conceptual model includes information on system boundaries; hydrostratigraphy; flow directions and sources and sinks; and a field-based estimate of water budget components. Uncertainty is addressed by updating and revising the conceptual model and/or formulating alternative conceptual models.

Chapter 6 – More on Sources and Sinks

Pumping and injection wells in the interior of a model are point sinks and sources that are typically represented by well nodes with assigned specified flow. Other sources and sinks in groundwater models include recharge from infiltration, leakage, underflow, evapotranspiration, springs, seeps, drains, and surface water features (streams, lakes, and wetlands). Some codes (e.g., MODFLOW) have packages to facilitate simulation of sources and sinks. However, other codes (e.g., FEFLOW) require the modeler to select an appropriate property or boundary condition from the options included in the code to represent a particular source or sink (e.g., a specified flow boundary condition to represent a pumping well).