Abstract

A catastrophic Palu earthquake on September 28th, 2018 with Mw 7.5 triggered countless slope failures, generating numerous landslides. This paper presents a practical method for susceptibility assessment of earthquake-induced landslides in the Palu region and the surrounding area. The statistical weight of evidence (WoE) model was used to assess the relationship between landslides induced by seismic motion and its causative factors to determine the susceptibility level and derive an earthquake-induced landslide susceptibility map of this study area. The 1273 landslides were classified into two data series, training data for modelling (70%) and test data for validation (30%). The six selected thematic maps as landslide causative factors are lithology, land use, peak ground acceleration (PGA), and slope (gradient, aspect, elevation). The selection of causative factors considerably influences the frequency of landslides in the area. The result is satisfactory because the AUC value of the chosen model excelled the minimum limit, which is 0.6 (60%). The estimated success rate of the model is 85.7%, which shows that the relevancy of the model is good with the occurrence of landslides. The prediction rate of 84.6% indicates that the applied model is very good at predicting new landslides.

Full Text
Published version (Free)

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call