Landslide hazard potential determination along a shoreline segment

Abstract Landslide hazard zoning is gaining importance in recent years. In such endeavours one deals with areas of large extent with limited data on material properties and certain other controlling factors. However, zoning is not a site-specific design activity and requires certain generalizations to be made. While field observations are essential, analytical procedures, where applicable, are of tremendous help in assessing the level of hazard potential. A probabilistic approach, used in assessing the slumping hazard of bluffs along the southwestern shoreline of Lake Superior is described. In this approach, those factors which are to be estimated, such as internal friction angle, cohesion intercept, position of stratigraphic unit interfaces, and ground-water table, based on limited actual data are generated in random combinations using the Monte Carlo method. Twenty conventional slip circle analyses using these sets of random combinations suffice to provide an assessment of landslide probability. Segments of the shoreline are classified on the basis of three probabilistic classes and the landslide probability thus assigned is compared with the hazard potential determined using the conventional deterministic method. The probabilistic approach is found to be a useful tool which minimizes the pressure to make arbitrary decisions while not eliminating engineering judgement and use of all existing information. There is some correspondence between the hazard classification given by the deterministic and the probabilistic methods; however, the correlation is not a good one. The probabilistic approach provides a more rational basis for determining the landslide hazard potential and for hazard zoning.