Control of glacial quarrying by bedrock joints

Abstract

A principal assumption of models that describe bedrock quarrying by glaciers is that cracks in the bed are small and isolated. These cracks are assumed to grow slowly in response to glacial loading to eventually delimit the areas of quarried surfaces. If this assumption is correct, then quarried-surface orientations will be controlled by orientations of principal stresses in the bed associated with sliding and resultant ice-bed separation downglacier from rock ledges and roches moutonnées, rather than by orientations of joints that predate glacial loading. To test this hypothesis, we compared orientations of quarried surfaces and preglacial joints in nine recently deglaciated forefields of glaciers in Canada and Switzerland with different bedrock lithologies and bedforms. In all but one forefield, quarried surfaces and major joint sets are coincident, with orientations of quarried surfaces bearing no systematic relationship to sliding direction other than their obvious tendency to not face upglacier. A difference in sliding direction of 64° between two areas of one glacier forefield had no effect on the orientation of quarried surfaces, with the same joint set exploited in both areas. These observations motivate a new characterization of subglacial bedrock that differs from that of existing quarrying models. The bed is idealized as a group of rock blocks separated by preglacial joints. Slow crack growth necessary for quarrying would be confined to small bodies of rock (bridges) that connect the faces of joints. This model of the bed does not allow it to be treated as an intact elastic solid with hydraulically isolated cracks as is usually assumed, which would have a profound influence on characterization of quarrying mechanics.