An Uncertainty Analysis of the Newmark Displacement Model for Earthquake-Induced Landslides in the Jiuzhaigou National Geopark

In this paper, we present an approach to generating probabilistic hazard maps for earthquake-induced landslides using the Newmark Displacement Model (NDM). This model takes the uncertainties associated with the slope properties (e.g., soil shear strengths, groundwater table location) into consideration, which is coupled with the hydrological model based on geomorphological, geological, geotechnical, seismological, and rainfall data. Uncertainties and fluctuations in the input parameters of the NDM are considered by treating these quantities as β-PERT distributions through Monte Carlo techniques, which allows probability value of the NDM to be cast into hazard maps. Additionally, incorporating Monte Carlo techniques can avoid using conservative input parameters in a deterministic approach to capture these uncertainties. Taking the 2017 Jiuzhaigou M_w 6.5 Earthquake in Sichuan Province, Western China as an example, earthquake-induced landslides probability distribution map is generated with the most appropriate displacement threshold (λ = 1 cm). Our results show good performances for realistic landslide hazard assessment, which can serve as a basis for providing a reference for the prediction of earthquake-induced landslide probability and rapid landslide hazard assessment after a strong earthquake.