Modelling of rockfall runout range : Employing empirical and dynamical methods

Abstract

Literature studies considering the determination of rockfall runout range have shown that the utilization of rigorous physical models yields significantly different results. This is primarily due to the fact that rockfall is a complex motion involving several modes of movement from free fall to bouncing, rolling and sliding. This has lead to difficulties in assigning the loss factor determined by the coefficient of restitution. Furthermore, investigations have shown that the restitution is dependent on many variables other than just material properties. The significant variance of variables and a limited knowledge of the dependence of each physical factor in relation with the retardation, make detailed numerical modeling of the phenomenon difficult and time consuming.

As an alternative we may utilize general terrain parameters common for all mountainsides to derive the runout distance. This thesis has focused on the statistical relations between the parameters, and how these can be used to derive an approximated runout range. In this respect two are based on simple linear regression results, while one is based on a multiple regression. In addition two simplified dynamic models have been put to the test. The results indicate that simple linear regression models may yield a relatively good first estimate of runout distance. However, the best statistical result was found with the use of a multiple regression model.

Keywords: rockfall, runout, dynamic modeling, empirical modeling, coefficient of restitution, dynamic friction coefficient, talus, topographical parameters, linear regression.