Assessing Landslide Susceptibility by Coupling Spatial Data Analysis and Logistic Model

Abstract

Landslides represent one of the most critical issues for landscape managers. They can cause injuries and loss of human life and damage properties and infrastructure. The spatial and temporal distribution of these detrimental events makes them almost unpredictable. Studies on landslide susceptibility assessment can significantly contribute to prioritizing critical risk zones. Further, landslide prevention and mitigation and the relative importance of the affecting drivers acquire even more significance in areas characterized by seismicity. This study aimed to investigate the relationship between a set of environmental variables and the occurrence of landslide events in an area of the Apulia Region (Italy). Logistic regression was applied to a landslide-prone area in the Apulia Region (Italy) to identify the main causative factors using a large dataset of environmental predictors (47). The results of this case study show that the logistic regression achieved a good performance, with an AUC (Area Under Curve) >70%. Therefore, the model developed would be a useful tool to define and assess areas for landslide occurrence and contribute to implementing risk mitigation strategy and land use policy.

Landslide susceptibility mapping using O-CURE and PAM clustering algorithms

“…In the past, most of the susceptibility maps were produced based on an expert's judgment which consumes a lot of time and energy, and it is difficult to quantify its accuracy due to its subjective effects. Fortunately, with the development of computer technologies such as Geographic Information Systems, Remote Sensing, and advanced data collection methods, machine learning algorithms are widely used in this field (Chowdhuri et al, 2021a;Ganga et al, 2022;Zhang et al, 2022). These developments have significantly improved LSM accuracy and efficiency.…”

“…Although most SL methods [examples: (Benchelha et al, 2020;Nhu et al, 2020;Arabameri et al, 2021;Berhane et al, 2021;Mehrabi and Moayedi, 2021;Ruidas et al, 2022b;Ganga et al, 2022;Korma, 2022;Sheng et al, 2022)] are popular and may achieve high prediction accuracy, they require data with predefined labels (landslide or non-landslide labels). Moreover, to obtain high performance these methods require a large number of labeled samples during the training process.…”

Analysis of landslide explicative factors and susceptibility mapping in an andean context: The case of Azuay province (Ecuador)

“…The authors pretend to answer the following questions based on the described scientific and practical issues: (i) How do LR and EFA evaluate the significance of landslide occurrence factors? LR has been widely used for landslide susceptibility mapping, as it allows identifying significant factors related to landslides [ [43] , [44] , [45] ]. However, EFA, as a multicriteria technique, has not been exploited for Landslide factor analysis as in other areas [ 46 , 47 ].…”

Landslide Susceptibility Evaluation Based on Potential Disaster Identification and Ensemble Learning

“…Thus, an imperious demand is suggested for the valid prevention and control of landslides to effectively mitigate the massive damage to people and engineering infrastructures. Landslide susceptibility evaluation (LSE) can predict the dangerous region where landslides may occur [ 1 ]; thus, it has become a crucial technique for disaster prevention and attracted great attention from the worldwide scientists [ 2 , 3 , 4 , 5 , 6 , 7 ]. Moreover, about 80% of these disastrous landslides have not been discovered in advance [ 8 ], called as potential or hidden landslides, which causes disaster control measures to miss the best time.…”