Review of mechanisms, methods, and theory for determining recharge to shallow aquifers in North Dakota
Effective management of ground-water resources requires knowledge of all components of the water budget for the aquifer of interest. Efforts to simulate ground-water flow prior to development and the effects of proposed pumping in several of North Dakota's shallow glacial aquifers have been hindered by the lack of reliable estimates of ground-water recharge. This study was done to (1) review the methods that have been used to measure recharge, (2) review the theory of unsaturated flow and the methods for characterizing the physical properties of unsaturated media, (3) consider the relative merits of a rigorous data-intensive approach versus an estimation approach to the study of recharge, and (4) review past and current agronomic research in North Dakota for applicability of the research and the data generated to the study of recharge. Direct, quantitative techniques for evaluating recharge are rarely applied. The theory for computing fluxes in unsaturated media is well established and numerous physics-based models that effectively implement the theory are available, but the data required for the models generally are lacking. Many parametric approaches have been developed to avoid the large data requirements of the physics-based approaches for analyzing flow in the unsaturated zone. However, the parametric approaches normally include fitting coefficients that must be calibrated for every study site, thereby detracting from the general utility of the parametric approach. The functional relation of matric potential to moisture content is required for physics-based soil-water models, whether analytic or numeric. Laboratory methods to determine these relations are tedious, costly, and may not give results representative of the soils as they occur in the field. Many models have been proposed to estimate the moisture-characteristic curve and hydraulic-conductivity function from basic soil properties, but none yield results that are universally satisfactory. In situ methods, because they require minimal disturbance of the soil profile and may be used repeatedly on the same soil mass, have become the preferred means for acquiring physical data, especially hydraulic conductivity. Hydro logic investigations, except for recent studies of hazardous-waste disposal sites, rarely have included physical characterizations of unsaturated media. Any of four phenomena could hinder attempts to simulate unsaturated flow in settings typical of North Dakota; variability of soil properties, hysteresis, frozen ground, and macropore development. The spatial and temporal variability of soil properties probably is the greatest complicating phenomenon and must be dealt with by detailed characterization of the properties. Hysteresis can detract from the accuracy of flow calculations for some soils under certain conditions but, for the present, our scant knowledge of soil physical properties is a greater hindrance to reliable soi1-water mode 1 ing than is the hysteresis phenomenon. A1 though seasona1ly frozen ground undoubtedly affects hydrologic processes in North Dakota, much more research is needed before meaningful quantitative treatment is possible. Finally, macropores can influence soil-water movement significantly, but macropore development may not be common on the intensively farmed, coarse-textured soils that typically overlie North Dakota's glacial aquifers. Lysimetry currently is the only reliable means of analyzing macropore flow. The soil-related research that has been conducted in North Dakota to date (1983) provides little of the type of information required to estimate ground-water recharge. Useful data could be developed by systematically evaluating the hydraulic characteristics of the prominent soil types overlying North Dakota's shallow glacial aquifers. These data would be required to enable use of a physics-based approach to estimating recharge. The size of the aquifer under study, its economic value, and the resources available for data collection should be considered when choosing between parametric or physics-based methods.