Mass movement susceptibility prediction and infrastructural risk assessment (IRA) using GIS-based Meta classification algorithms

Abstract

In mountainous areas, mass movements are among the most dangerous natural hazards. Infrastructure is a crucial component and is thought of as human wealth. This infrastructure is frequently impacted by mass movements, whose frequency and size are anticipated to rise in the future due to the unequal distribution of rainfall events brought on by climate change. To deal with the anticipated repercussions, the study area, the Northern part of Morocco, needs to implement new management and maintenance practices. Thus, the main motivation of this study was to examine the mass movement vulnerability and assess risk on infrastructures in two provinces of North Morocco, i.e. Chefchaouen and Tetouan. The present study employed Reduced Error Pruning Tree (REPTree) and its ensemble with Bagging, AdaBoost, and Random SubSpace (i.e. REPTreeBagging, REPTreeAdaBoost, and REPTreeRandomSubSpace) for mass movement susceptibility mapping (MMSM) based on a comprehensive dataset of 100 mass movements locations which include debris flow, landslide, and rock fall during past 20 years (2000–2020) as well as 12 MM conditioning factors. The result revealed that REPTreeRandomSubSpace is the most viable model for MMSM with AUC = 0.8656. In addition, REPTreeBagging, REPTreeAdaBoost, and REPTree models also offer acceptable results with AUC of 0.8338, 0.8269, and 0.7942, respectively. After MMSM, the infrastructural risk was assessed and the result showed that among the six infrastructural features considered, buildings and forests have a greater risk of mass movement in the study area. Most of the mass movement and infrastructural risk-prone areas are found in the central and north-eastern parts of the study area. The results further revealed that elevation, land use, lithology, rainfall, and distance from roads are important variables for mass movement and infrastructure risk assessment (IRA). The results of this study offer a systematic sight for decision-makers to mitigate natural disasters and infrastructural risk in the study region.