Some constraints on the severity of landslide penetration in sensitive deposits

Theoretical considerations and case-studies are presented to show that, in many cases, the severity of retrogression of an unstable valley slope, in areas of sensitive muddy sediment, is controlled by the topography of the valley. A formula is offered to predict the distance of retrogression from topographic attributes of valleys. It is suggested that most retrogressive landslides in sensitive sediments involve only limited liquefaction of the spoil, and it is for this reason that retrogression is controlled by this topographic constraint. Those situations in which retrogression stops before this limit is reached are also discussed: one important factor which can determine whether or not such aborted retrogression will occur appears to be the nature of the first-time slide. Those situations in which retrogression can exceed this topographic limit are briefly examined as well: attention is focussed on the importance of spoil liquefaction as a prerequisite for such excess landslide retrogression. Data are presented which indicate that the initial undrained strength of the sediment exerts a major control on the degree of spoil liquefaction. Finally the possibility is considered that some assumed retrogressive failures were in fact flake slides in which the slide mass disintegrated after failure. Such 'retrogressive facsimiles' are considered to be rare.

On the retrogression of landslides in sensitive muddy sediments

The literature dealing with extensive retrogression of landslides in sensitive sediments is reviewed and found to be inadequate in several respects. One obvious deficiency is in the link between supposed retrogressive mechanisms and the morphology of the resultant ‘earthflow’ cavity. Detailed attention is focussed on the origin of linear clay ridges that protrude from the floors of fresh earthflows, and persist for only a short time before being degraded to micro-scale features. The theory of landslide retrogression proposed by Odenstad, after the Sköttorp landslide in Sweden, is discussed, and developed into a model for the prediction of landslide retrogression, and for the explanation of 'flowbowl' morphology. The model, essentially an undrained retrogressive spreading failure, is tested against selected retrogressive landslide sites for which all the necessary information is available; the agreement between nature and theory is good. Some speculation is made regarding the origin of bottlenecked flowbowls, which are viewed as the exception rather than the rule in the sensitive muds of the St. Lawrence Lowlands. Emphasis is directed at the importance of the undrained shear strength of sensitive sediments, and less significance is attached to the actual magnitude of sensitivity of such deposits. Finally, the limitations of the model in very soft sediments, such as found in parts of Norway, are briefly pointed out.