Obtaining suitable logic-tree weights for probabilistic earthquake-induced landslide hazard analyses

In this study, an inventory of landslides induced by the 2011 Lorca earthquake (Mw 5.1) has been used in order to develop a new procedure to obtain objective logic-tree weights for a probabilistic earthquake-induced landslide hazard analysis. The 2011 Lorca earthquake triggered more than 250 landslides, mainly of disrupted type. The logic-tree was designed having regard to variability of relevant geotechnical parameters involved in the problem and uncertainties associated with the use of several empirical relationships in order to compute Newmark displacements. For the purpose, the resulting hazard maps were compared with this landslide inventory, and weights estimated for each branch of the logic tree based on these results. The best model for seismic landslide hazard mapping for a moderate earthquake correctly identifies around 72% of landslide areas. Based on the set of parameters that comprises (depth of failure surface, specific weight, cohesion, friction angle and Newmark displacement model), the corresponding weights were objectively established. These weights are reliable enough for the obtaining seismic landslide hazard maps and may be implemented in similar environments characterized by moderate-low magnitude earthquakes (Mw < 5.5).