Surface time-domain EM sounding was completed at three sites at the
Atomic Energy of Canada Limited (AECL) Underground Research Laboratory
(URL), Lac du Bonnet, Manitoba to investigate the feasibility of delineating
the interface between near-surface fresh groundwater and deeper saline
water. The interface occurs at a depth of 150 to 300 m within the resistive
granite rock of the Lac du Bonnet batholith. Recordings were made using
a PROTEM47 system in order to provide a sufficiently early time-domain
response to detect the interface. Measurements of the background EM noise
were completed at the site. The noise was found to decrease by a factor
of five from locations near powerlines to more distant locations.
Time-domain EM soundings were completed using square 10, 20 and 40 m transmitter loops at Site 1; 40 m transmitter loop in Site 3. At Site 1, the results for the smaller transmitter loops were erratic due to a spatially variable conductive layer near the surface. Results for the 40 m loop were consistent for offsets ranging from zero to about 25 m. We use 40 m central- loop response to determine the conductivity structure. Forward modelling was completed to determine the general form of the conductivity model and geophysical inversion was then completed. The results indicate a three-layer structure. The first layer is a thin conductive layer (<10 m), the middle is very resistive and the lower layer is more conductive. Constrained inversions indicate that the depth to the conductive layer is 195+/-20 m and that the ratio of conductivity of the two layers exceeds 1:80. At Site 3, the results show that the top layer is more conductive and thicker. The 40 m transmitter loop is not large enough to determine the deep structure because of the effects of this top layer. A 100 m transmitter loop is recommended. Comparison of the EM results with nearby borehole logs in Site 1 provides a strong indication that the conductivity interface corresponds to the fresh/saline-water interface. The results of the survey thus suggest that the surface time-domain EM method can be used to successfully map the groundwater interface. Further modelling of the results is continuing in order to assess the optimal configuration of an EM survey to map the interface.