Gelifluction: Observations from Large-Scale Laboratory Simulations

Abstract

Despite extensive field studies, progress in understanding gelifluction processes has been limited. Controlled laboratory simulation experiments offer an alternative and potentially extremely effective approach. Such an experiment is described here. It was conducted on a 12° slope formed of two natural soils, one a fine sandy silt derived from slate bedrock, the second a gravelly silty sand derived from mudstone bedrock. Continuous measurements were made of soil temperatures, porewater pressures, frost heave, thaw settlement, and downslope displacements of the soil surface over seven freeze/thaw cycles. Two-dimensional vectors of soil surface movements together with evidence from excavated displacement columns suggest that gelifluction occurred only during thaw consolidation of the upper parts of the soil profile; thawing of the deeper layers caused thaw consolidation but little downslope displacement. Cryogenic processes are shown to cause progressive decreases with depth in void ratio and moisture content and increases in undrained shear strength within the continuous soil matrix that separates ice lenses. Since self-weight stress levels are low, thawing leads to significant shear strain only in the softer, wetter near-surface soil layers.