The geotechnical investigation and stabilization design for a landslide located near Fairbanks, Alaska, revealed several issues with testing and modelling landslide shear zones in permafrost materials. The subsurface soils consist of frozen silts with zones of organics and ice inclusions (wedges and lenses), and underlain by zones of unfrozen sand and gravel. A 30 m deep mining pit was excavated downslope of a highway, effectively removing lateral support and causing instability. Borings were drilled and instrumented with inclinometers, standpipe observation wells, vibrating-wire piezometers and thermistors. Measured ground movements indicated that the shear zone passed through both frozen and unfrozen soils. The literature was researched to check for documentation of residual shear strengths on frozen silts; however, there is very little information published on this subject. Attempts to back-calculate the effective residual friction angles were complicated since the actual groundwater pressure distribution along the shear zone could not be reliably estimated. Ultimately, the design of stabilization measures was based on total stress analysis. Back-calculated values of residual strength were expressed in terms of “average cohesion.” This analysis method assumes that groundwater conditions after stabilization would be similar to the levels at the time of failure.
