A Ground Penetrating Radar Study of Water Table Elevation in a Portion of Wisconsin’s Central Sand Plain

Abstract: Contamination of groundwater by agricultural chemicals in the Central Sand Plain has prompted studies of groundwater flow in this region. Because the groundwater system is particularly susceptible to contamination in areas where groundwater recharge occurs, identification of recharge zones can contribute significantly to the effective management of agricultural chemical use. An accurate map of water table elevation (groundwater head) is crucial to identifying the distribution of recharge.

This project tests the reliability of ground penetrating radar (GPR) as a tool for obtaining high-resolution maps of water table elevation. GPR surveys were performed in the summer and fall of 1987 in an approximately nine square mile area in the Central Sand Plain with a fairly dense set of existing water table observation wells. GPR is a geophysical instrument which transmits an electromagnetic pulse into the ground and records the return times of pulses reflected from subsurface interfaces, producing a profile of return time versus horizontal distance as the radar is pulled along the ground. In order to test the reliability of GPR as a water table reconnaissance tool, sparse subsets of wells in the area are used to calibrate the radar and water table depths obtained from these calibrations are compared to known water table depths in the remaining wells. Thus, the study tests whether GPR can be used to help produce an accurate water table map when only a few observation wells exist.

Three wells are the minimum necessary to obtain an estimate of uncertainty in calibration parameters, specifically the radar signal velocity in the subsurface materials and the return time correction factor. Results indicate that, in the study area, three calibrating wells are adequate to give a correlation of 0.99 and a root mean squared deviation of about 1 foot between radar-predicted and observed water table depths. If several wells distributed throughout a region of interest do yield consistent calibration results, radar can be used to produce a map of water table elevation in that region. Attempts to model the distribution of groundwater recharge and discharge in the study area were not very successful. Model results appear to be highly dependent on the technique employed to interpolate a water table map from either the well or radar data. However, a water table map produced from radar data in an area with sparse well coverage would still be useful for guiding further hydrogeological work in that area.