Recent investigations suggest that dispersion in aquifers is scale dependent and a function of the heterogeneity of aquifer materials. Theoretical stochastic studies indicate that determining hydraulic-conductivity variability in three dimensions is important in analyzing the dispersion process. Even though field methods are available to approximate hydraulic conductivity in three dimensions, the methods are not generally used because of high cost of field equipment and because measurement and analysis techniques are cumbersome and time consuming.

The hypothesis of this study is that field-determined values of dispersivity are scale dependent and that they may be described as a function of hydraulic conductivity in three dimensions. The objectives of the study at the Bemidji research site are to (1) determine hydraulic conductivity of the porous media in three dimensions, (2) determine field values of dispersivity and its scale dependence on hydraulic conductivity, and (3) develop and apply a computerized data-collection, storage, and analysis system for field use in comprehensive determination of hydraulic conductivity and dispersivity.

Plans for this investigation involve a variety of methods of analysis. Hydraulic conductivity will be determined separately in the horizontal and vertical planes of the hydraulic-conductivity ellipsoid. Field values of dispersivity will be determined by single-well and doublet-well injection or withdrawal tests with tracers. A computerized data-collection, storage, and analysis system to measure pressure, flow rate, tracer concentrations, and temperature will be designed for field testing. Real-time computer programs will be used to analyze field data. The initial methods of analysis will be utilized to meet the objectives of the study.

Preliminary field data indicate the aquifer underlying the Bemidji site is vertically heterogeneous, cross-bedded outwash. Preliminary analysis of the flow field around a hypothetical doublet-well tracer test indicates that the location of the wells can affect the field value of dispersivity. Preliminary analysis also indicates that different values of dispersivity may result from anisotropic conditions in tests in which observation wells are located at equal radial distances from either the injection or withdrawal well.