An integrative approach of the physical-based stability index mapping with the maximum entropy stochastic model for risk analysis of mass movements

Abstract

This research was mainly aimed at the validation analysis of an integrative approach of the physical-based stability index mapping (SINMAP) with the maximum entropy (MaxEnt) stochastic model for risk analysis of mass movements to identify effectual driving forces. The study area (the geologic zone of the Kopet Dagh-Hezar Masjed) is geographically located in the northeast of Iran, where mass movements had been recorded in the types of slide, fall, compound zone and mudflow. Different layers of information including topography, geology, land use and vegetation, rainfall, and soil properties were extracted and analyzed in the geographic information system. The effective factors on the incidence of each mass movement group were determined based on the results of the Jackknife estimator. The approach of receiver operating characteristic was used to validate the MaxEnt results. According to the validation data set, the area under the curve for the incidence modeling of the slide, fall, compound zone, and mudflow was 0.723, 0.749, 0.729, and 0.727, respectively, which demonstrated good predictions by the SINMAP-MaxEnt hybrid model. The integration of SINMAP with MaxEnt was able to improve the results (up to 3–8%), as compared with employing only MaxEnt, through estimating the hydrological and geotechnical parameters especially in places where we faced a shortage of observational records. Stability index analysis showed that 50% of all recorded points of mass movements were to be found in a naturally stable zone. This examination accompanied with sensitivity analysis demonstrated that significant destabilizing factors, especially anthropogenic driving forces like land-use change have led to slope instability.