Operational estimation of landslide runout: comparison of empirical and numerical methodologies.

Abstract:

A key point of landslide hazard assessment is the estimation of their runout. Empirical relations linking angle of reach to volume can be used relatively easily, but they are generally associated to large uncertainties. Indeed, they are often derived from databases compiling numerous landslides’ characteristics, without considering the topographic specificity of each case. On the contrary, numerical simulations provide more detailed results on the deposits morphology, but their rheological parameters can be difficult to constrain. Simulating all possible values can be time consuming and incompatible with operational requirements for rapid prediction.

We propose and compare three operational methods to derive scaling laws relating the landslide travel distance to the destabilized volume. The first one relies only on empirical relations. The other two use the shallow-water numerical code Shaltop with the Coulomb rheology to generate a database of multiple simulation runs. We deduce from this database a scaling law relating the travel distance to the initial volume and the friction coefficient. To get a law that depends on the volume only, the friction coefficient is then either deduced from previous events back-analysis, or directly from empirical relations. We analyze the propagation of uncertainty and compare the prediction of the final runout/volume scaling laws to observations.This methodology is applied to three case studies : the Fei Tsui road landslide in Hong Kong (14,000 m³, 100 m runout), the Samperre cliff collapses in Martinique, Lesser Antilles (0.5 to 4 Mm³, runout about 2 km) and the Frank Slide rock avalanche (36 Mm³, 3 km runout).

The higher estimation precision together with the lower width of the prediction interval is obtained with the Shaltop-derived scaling law, with the friction coefficient derived from back-analysis. This is true in particular for small landslides where the initial geometry strongly impact runouts, which is not well described by classical mobility indicators such as the angle of reach. Furthermore, numerical simulations take into account the complexity of the landslide path and bottom topography, which is hardly done with empirical relations. Finally, simulations also allow to distinguish between uncertainties related to volume and to landslide mobility.