Electromagnetic resistivity anisotropy in till from the Kiskatinaw streamlined landform field, northeastern British Columbia, Canada

Abstract

Applications of the azimuthal electromagnetic (EM) method that establish the internal structure of landforms consisting of sediment are sparse. In the Kiskatinaw streamlined landform field of northeastern British Columbia, ice-flow parallel till ridges formed beneath the Cordilleran Ice Sheet during the late Wisconsinan. EM anisotropy data from the Kiskatinaw field are systematically related to ridge elongation and orientation and, in part, to clast fabrics. However, in contrast to previous studies, we find that resistivity measurements do not necessarily record the orientation of ice flow, and electrical anisotropy in till may not be a simple indicator of ice-flow direction. For highly elongate ridges (l/w > 3), the long axes of resistivity ellipses are at high angles to the long axes of the landforms. Tills in these elongate ridges are matrix rich (>85%) and display strong unimodal ice-flow parallel clast fabrics. For more equant ridges (l/w < 3), the long axes of resistivity ellipses are at low angles to the long axes of the landforms. Tills in these ridges are clast rich (up to 60%) and display polymodal clast fabrics with both ice-flow parallel and ice-flow transverse modes. Because clast fabrics in the more equant ridges are so variable, despite unimodal resistivity patterns, and because groundwater well data indicate relatively fresh pore waters and the till matrix contains clay (average 12%), we speculate that the resistivity/conductivity anisotropy we observed may record aligned matrix clays rather than clast fabrics alone. We interpret that the highly elongate ridges indicate subglacial shear along a non-cohesive easily deformed bed, whereas the more equant ridges mark ‘sticky spots’ where high clast concentrations produced local areas that resisted subglacial shear and induced local ice-flow parallel shortening, which limited landform elongation and clast alignment.