The mapping assessment of rock fall probability is mostly based on qualitative observations. For this reason, expertise is of great importance in determining to which hazard class a sub-area belongs. Consequently, the resulting hazard map may depend on the person in charge of the field study. We present an experiment that was realized in order to evaluate the importance of subjective assessment: three populations, with different level of expertise (students in geosciences, researchers in geosciences and confirmed experts) have been asked to assess rock fall hazard on a test site divided in three different sectors. Their first assessment of the probability of a rock fall was given without using any method. Then, they used the "Laboratoire des Ponts et Chaussées" (LPC) method, which is a qualitative method frequently used in France and the "Slope Mass Rating" (SMR) method which is a more quantitative approach, resulting from the classifications of rock masses. An analysis of results variance shows that there is a significant influence of the method used on the assessment of rock fall probability, but a non-significant influence of the expertise level of the population. The SMR method seems highly sensitive to the presumed instable volume and exhibits an important dispersion on the results. To the contrary, with the LPC qualitative method, the results are similar.
Rock fall instabilities are a major risk for population, human activities and infrastructure (Bell & Glade 2004, Moreiras 2006). It is thus necessary to evaluate the level of rock fall hazard and propose hazard maps in order to manage risk in urban areas. Construction in some high hazard areas is excluded, or requires geotechnical engineering assessment of slope stability (Fell et al. 2008). In some parts of the world, landslide hazard and risk maps are introduced across the country. They provide different hazard classes as for example the PPR ("Plans de Prevention des Risques Naturels Previsibles") in France (Laboratoire des Ponts et Chaussées 1999) and the "Cartes de Dangers" in Switzerland (Leroi et al. 2005). This evaluation will also allow to establish a monitoring system of rock mass movement or a suitable protection. It must be as accurate as possible (not underestimated neither overestimated) and reproducible.
Rock fall hazard can be defined as the probability that a specific location at the toe of a studied slope is reached by a rock fall of a given intensity (Jaboyedoff et al. 2001). What ever is the intensity of the expected rock fall, the previous probability can be divided in two terms as (Hantz 2011): the probability of detachment (or probability of failure) and the propagation probability.