An historical, geomechanical and probabilistic approach to rock-fall hazard assessment

A new method (HGP for Historical, Geomechanical and Probabilistic) is proposed to estimate the failure probability of potentially unstable rock masses in a homogenous area, as a function of time. Analysis of a rock falls inventory yields the mean number of rock falls which may be expected in the area for the given time period and a given volume range. According to their geomechanical features, the potentially unstable rock masses are distributed in classes corresponding to different failure probabilities. The expected number of rock falls can be expressed as a function of these unknown probabilities. Assuming that only the ratio between these probabilities can be estimated, combining the historical and geomechanical analysis allows estimating the order of magnitude of the different failure probabilities. The method gives a quantitative significance to the evaluations which are usually attributed to potentially unstable rock masses. Rock-fall hazard can then be compared with other natural hazards, such as floods or earthquakes. The method is applied to a case study of calcareous cliffs in the area of Grenoble, France.