An Approach to Groundwater Monitoring Well Network Design Relying on Numerical Techniques and Public-Domain Information

Abstract

Uncertainties in the hydrogeology of a study area and in the transport properties of potential contaminants challenge the designers of groundwater monitoring well networks- Numerical flow modeling is a useful tool for guiding the placement of wells, but it requires justified boundary conditions and sufficient knowledge of aquifer parameters. In a graciated terrain, additional problems arise because of complex spatial arrangements of aquifers and aquitards. The transport of conservation contaminant tracers is normally calculated as a simple function of average values of hydraulic conductivity, hydraulic gradient, and effective porosity. But are the results of this straightforward method always valid? In order to address these topics, this investigation focused on a landfill on the Anoka sandplain of east-central Minnesota. The purpose of this study was to determine the proper placement of an initial group of monitoring wells at the landfill using only offsite public-domain data. The results may then be applied to the siting of other wells. Finite-difference flow modeling was supported by an abundance of inexpensive public-domain information and by the construction of a detailed, sub-regional glacial geologic map. A two-dimensional kriging analysis refined the model by determining the cell-by-cell best estimates of the basal elevation of the surficial aquifer. Particle tracking results indicated the expected pathway of landfill leachate. Based on the results, one well upgradient of the landfill and several downgradient wells were selected from the database of actual monitoring wells, and the head data from these shallow wells were used to calibrate the model. The calibrated hydraulic conductivity of the sandplain aquifer agrees closely with values obtained through grain-size analyses and pump tests. Numerical analyses of boundary conditions support the validity of the flow model. Other case studies of unconfined outwash aquifers suggest that predicted plumes of conservative tracers are often greater than the actual extents. Compared to the chloride data for monitoring wells at the Anoka site, particle tracking results have an accurate orientation but a length at least two times too long. Uncertainties, such as the effect of longitudinal dispersion and the transient nature of the leachate's initial concentration and source area, suggest an even greater difference' This conceptual understanding of plume migration provides guidance for the placement of additional downgradient wells. The described application of models and inexpensive offsite data to monitoring well network design is a methodology that may be effective for the monitoring of solutes from existing or proposed potential contamination sources.