Method of Up-Slope Mitigation Priority for Alishan Mountain Road in Taiwan

Abstract

Rainfall-induced up-slope failure has been a major issue for the mountain road mitigation. To determine the mitigation priority of the roadside slopes, a model to evaluate the stability of the up-slopes is proposed in this paper. This model is developed using Bayesian classification theorem, historical rainfall and landslide data, and an artificial intelligence skill—Gaussian process. The parameters needed for this model include the up-slope features, catchment area, lithology, vegetation coverage, and rainfall data. Among these parameters, rainfall is considered to be the triggering factor of landslide in this study. In addition, each slope has also demonstrated a unique failure probability relationship with the 72-h accumulated rainfalls. Such a relationship is adopted to define a threshold rainfall factor which is used as an index to estimate the likelihood of landslide occurrence under a rainfall event. Using the threshold rainfall factor of each slope, a priority list is proposed for the mitigation of roadside up-slopes along the Alishan mountain road in Taiwan.